US 7220838 B2
The present invention provides two novel polypeptides, referred to as the “N” and “C” fragments of hedgehog, or N-terminal and C-terminal fragments, respectively, which are derived after specific cleavage at a G↓CF site recognized by the autoproteolytic domain in the native protein. Also included are sterol-modified hedgehog polypeptides and functional fragments thereof. Methods of identifying compositions which affect hedgehog activity based on inhibition of cholesterol modification of hedgehog protein are described.
1. An isolated antibody which binds specifically to the N-terminal fragment of a human sonic hedgehog polypeptide, wherein the N-terminal fragment has a Glycine residue at its carboxy terminus, resulting from specific cleavage of a G⇓CF cleavage site specifically recognized by a proteolytic activity of the carboxy terminal fragment of a native human sonic hedgehog polypeptide.
2. The antibody of
3. The antibody of
4. The antibody of
5. The antibody of
6. The antibody of
7. The antibody of
8. A method of generating an antibody which binds specifically to the N-terminal fragment of a human sonic hedgehog polypeptide, comprising immunizing a mammalian subject with the N-terminal fragment, wherein the fragment has at its carboxy terminus, a Glycine residue resulting from specific cleavage of a G⇓CF cleavage site specifically recognized by a proteolytic activity of the carboxy terminal fragment of a native human sonic hedgehog polypeptide.
9. The method of
10. The method of
11. The method of
12. The method of
This application is a divisional and claims the benefit of priority under 35 USC § 121 of U.S. application Ser. No. 09/562,914, filed May 2, 2000, issued Jun. 28, 2005 as U.S. Pat. No. 6,911,528; which is a divisional application of U.S. application Ser. No. 08/946,329, filed Oct. 7, 1997, issued May 2, 2000 as U.S. Pat. No. 6.057,091, which is a continuation-in-part of U.S. patent application Ser. No. 08/729,743, filed Oct. 7, 1996, issued Apr. 10, 2001 as U.S. Pat. No. 6,214,794, which is a continuation-in-part of U.S. patent application Ser. No. 08/567,357, filed Dec. 4, 1995, issued Oct. 17, 2000 as U.S. Pat. No. 6,132,728, which is a continuation-in-part of U.S. patent application Ser. No. 08/349,498, filed Dec. 2, 1994, issued Aug. 28, 2001 as U.S. Pat. No. 6,281,332. U.S. application Ser. No. 08/946,329, filed Oct. 7, 1997, issued May 2, 2000 as U.S. Pat. No. 6,057,091, also claims priority under 35 USC § 119(e) to U.S. patent application Ser. No. 60/061,323, filed Oct. 2, 1997, now abandoned. The disclosure of the prior applications are considered part of and are incorporated by reference in the disclosure of this application.
1. Field of the Invention
This invention relates generally to the field of protein processing and protein signalling pathways and specifically to two novel proteins having distinct activities, which are derived from a common hedgehog protein precursor.
2. Description of the Related Art
Embryologists have long performed experimental manipulations that reveal the striking abilities of certain structures in vertebrate embryos to impose pattern upon surrounding tissues. Speculation on the mechanisms underlying these patterning effects usually centers on the secretion of signaling molecule that elicits an appropriate response from the tissues begin patterned. More recent work aimed at the identification of such signaling molecules implicates secreted proteins encoded by individual members of a small number of gene families. One such family of proteins which may have an influential effect upon patterning activities are those proteins encoded by the hedgehog gene family.
The hedgehog (hh) gene was initially identified based on its requirement for normal segmental patterning in Drosophila (Nũsslein-Volhard, C. & Wieschaus, E, Nature 287:795–801, 1980). Its functions include local signaling to coordinate the identities of adjacent cells within early embryonic segments (Hooper, J. E., & Scott, M. P. Early Embryonic Development of Animals, pp. 1–48, 1992) and a later function in cuticle patterning that extends across many cell diameters (Heernskerk, J. & DiNardo, S., Cell, 76:449–460, 1994). The hh gene also functions in the patterning of imaginal precursors of adult structures, including the appendages and the eye (Mohler, J. Genetics, 120:1061–1072, 1988; Ma, et al., Cell, 75:927–938, 1993; Heberlein, et al., Cell, 75:913–926, 1993; Tabata, T. & Komberg, T. D., Cell, 76:89–102, 1992; Basler, K. & Struhl, G., Nature, 368:208–214, 1994). Genetic and molecular evidence indicates that hedgehog proteins are secreted and function in extracellular signaling (Mohler, J., supra; Lee, et al., Cell, 711:33–50, 1992; Taylor, et al., Mech. Dev., 42:89–96, 1993).
In vertebrates activities encoded by hh homologues have been implicated in anterior/posterior patterning of the limb (Riddle, et al., Cell, 75:1401–1416, 1993; Chang, et al., Development, 120:3339, 1994), and in dorsal/ventral patterning of the neural tube (Echelard, et al., Cell, 75:1417–1430, 1993; Krauss, et al., Cell, 75:1431–1444, 1993; Roelink, et al., Cell, 76:761–775, 1994).
The vertebrate ventral midbrain contains neurons whose degeneration or abnormal function are linked to a number of diseases, including Parkinson's disease and schizophrenia. It is known that motor neurons develop in close proximity to the floor plate in the ventral midbrain. Midbrain projections to the striatum are involved in the control of voluntary movement (Bjorklund and Lindvall, In: Handbook of Chemical Neuroanatomy, eds., Borklund, et al., Amsterdam: Elsevier, pp55–122, 1984) and loss of these neurons results in the motor disorders of Parkinson's disease (Hirsch, et al., Nature, 334:345, 1988). Midbrain dopaminergic neurons that innervate limbic structures and the cortex influence emotional and cognitive behavior, respectively, and abnormal function of these neurons has been associated with schizophrenia and drug addiction (Seeman, et al., Nature, 365:441, 1993).
While the molecular nature of the factors that specify neuronal cell fate have not been established, members of the transforming growth factor-β (TGF-β) (Lyons, et al., Trends in Genetics, 7:408, 1991) or the hedgehog protein family (Smith, J. C., Cell, 76:193, 1994) may possess the characteristics expected from such factors as they participate in specification of cell fate, mediate inductive interactions between tissues, and in many cases act at a distance of only a few cell diameters.
The present invention establishes that hh activities encoded by these genes play a crucial role in early patterning of the developing eye and in patterning of the brain. For the first time, the invention shows that internal cleavage of hedgehog protein product is critical for full function, and that the two novel products of this auto-proteolytic cleavage display distinguishable activities, thus demonstrating that hh signaling activity is a composite effect of two separate signaling proteins that derive from a common hh protein precursor. In so doing, the invention provides the means for specific patterning and proliferation of desired neuronal cell types for addressing disorders which arise from neuronal degeneration or abnormal function.
The present invention is based on the seminal discovery that hedgehog proteins undergo auto-proteolytic cleavage which results in two separate proteins having distinct functional and structural characteristics. The two polypeptides, referred to as the “N” and “C” fragments of hedgehog, or N-terminal and C-terminal fragments, respectively, are produced after specific cleavage at a G⇓CF site recognized by the autoproteolytic domain in the native protein. The “C” fragment functions as a cholesterol transferase during autoproteolysis thus allowing cholesterol modification of the “N” fragment.
Thus, in one embodiment, the invention provides a substantially pure polypeptide characterized by having an amino acid sequence derived from amino terminal amino acids of a hedgehog protein and having at its carboxy terminus, a G⇓CF cleavage site specifically recognized by a proteolytic activity of the carboxy terminal fragment of the native hedgehog polypeptide. The invention also provides a substantially pure polypeptide characterized by having an amino acid sequence of a hedgehog polypeptide or a fragment derived from amino terminal amino acids of a hedgehog polypeptide, wherein the polypeptide or fragment thereof comprises a sterol moiety. Fragments derived from a native hedgehog polypeptide are included and preferably include extracellular amino acid residues, such as those derived from the N fragment. In one embodiment of the invention, the sterol moiety is cholesterol.
In another embodiment, the invention provides a substantially pure polypeptide characterized by having an amino acid sequence derived from carboxy terminal amino acids of a hedgehog protein and having at its amino terminus, a G⇓CF cleavage site specifically recognized by a proteolytic activity of the carboxy terminal fragment of the native hedgehog polypeptide.
The invention also provides a method for modulating proliferation or differentiation of neuronal cells, comprising contacting the cells with a hedgehog polypeptide. The native hedgehog polypeptide, the N, or the C fragment, or functional fragments derived therefrom, are most useful for the induction of proliferation or differentiation of neuronal cells substantially derived from floor plate neuronal cells.
In yet another embodiment, the invention provides a method for identifying a compound which affects hedgehog activity comprising incubating the compound with hedgehog polypeptide, or with biologically active fragments thereof, or with a recombinant cell expressing hedgehog, under conditions sufficient to allow the components to interact; and determining the effect of the compound on hedgehog activity or expression. For example, cholesterol level (e.g., biosynthesis or transport) is measured as an indicator of hedgehog activity. In one aspect of the invention, the method provides a means for affecting cholesterol biosynthesis or transport in a cell comprising contacting a cell with an effective amount of a compound that affects hedgehog, thereby affecting cholesterol biosynthesis or transport. The effect may be inhibition or stimulation of cholesterol biosynthesis or transport.
FIGS. 28B-1–28B-3 are topology diagrams of Hh-C17 (SEQ ID NO:46 . Residues in β strands are in boxes with amino-acid type and number indicated. Residues in turns of 310 helix are ovals with amino-acid type and number indicated. Other residues in the structure are in boxes with amino-acid number indicated. Hydrogen bonds between β strands are indicated with arrows. A pseudo two-fold axis of symmetry is indicated with a diamond. This panel was prepared using the output of the program PROMOTIF.
The present invention provides two novel polypeptides originally derived from a single precursor protein, both of which have distinct structural and functional characteristics. The proteins are derived from a hedgehog protein and can be naturally produced by auto-proteolytic cleavage of the full-length hedgehog protein. Based on evidence provided herein, which indicates that hedgehog precursor protein and the auto-proteolytic products of hedgehog precursor protein are expressed in the floorplate of the ventral midline of the neural tube and notochord, the invention now provides a method for the induction of proliferation or differentiation of neuronal cells associated with or in close proximity to the Doorplate and notochord. The invention also provides cholesterol modified hedgehog polypeptides and function fragments thereof.
In a first embodiment, the invention provides a substantially pure polypeptide characterized by having an amino acid sequence derived from amino terminal amino acids of a hedgehog protein and having at its carboxy terminus, a glycine-cysteine-phenylalanine (G⇓CF) cleavage site specifically recognized by a proteolytic activity of the carboxy terminal fragment of the native hedgehog polypeptide. This fragment is denoted the N-terminal fragment or polypeptide or “N”, herein. For example, in the case of the Drosophila hedgehog, the N fragment includes amino acids 1–257 of hedgehog protein, wherein amino acids 85–257 have a molecular weight of about 19 kD by non-reducing SDS-PAGE (Amino acid residue numbers 1–257 include non-structural features such as signal sequences.). The G⇓CF cleavage site in Drosophila hedgehog precursor protein occurs at amino acid residues 257–259. Those of skill in the art will be able to identify the G⇓CF cleavage site in other hedgehog genes, as the amino acid location will be similar and the site will be specifically recognized by the autoproteolytic activity of the corresponding C fragment.
The N-terminal polypeptide is also characterized by being cell-associated in cells expressing the polypeptide in vitro, and being specifically localized in vertebrate or Drosophila cells or embryos, for example. In other words, this N-terminal fragment of hedgehog, remains close to the site of cellular synthesis. The association of N with the cell is a result of the processing event which involves lipophilic modification of the amino terminal domain. (See
The N polypeptide of the invention is characterized by having an amino acid sequence derived from amino terminal amino acids of hedgehog protein, e.g., 1–257 in Drosophila, wherein amino acids 1–257 have a molecular weight of about 19 kD by non-reducing SDS-PAGE. The N polypeptide includes smaller fragments which retain the functional characteristics of full length N, e.g., bind to heparin. The hedgehog protein from which N is derived includes, but is not limited to Drosophila, Xenopus, chicken, zebrafish, mouse, and human. Crystallographic analysis shows the structure of SHH-N includes the presence of a zinc ion. While not wanting to be bound by a particular theory, the presence of the zinc ion is suggestive of zinc hydrolase activity. Zinc hydrolases include proteases such as carboxypeptidase A and thermolysin, lipases such as phospholipase C, and other enzymes such as carbonic anhydrase. Alterations in the zinc hydrolase site of the amino terminal signaling domain may be useful for modulating the range of diffusion of a hedgehog protein or to alter the signaling characteristics of the amino terminal signaling domain. For example, a mutation in the zinc hydrolase site may result in a tethered protein where ordinarily the protein is secreted at a distance. The result would be induction of a cell type not typically induced. Alteration in the zinc site may result in a molecule capable of inducing motor neurons and not floor plate, and vice versa.
The identification of a cell-surface, or extracellular matrix localization of N and its expression in notochord and floor plate-associated cells, provides a means for isolation or specific selection of cells expressing N, e.g., to isolate a notochord sample or to isolate floor plate cells. In addition, antibodies directed to N are useful for histological analysis of tissues suspected of expressing N protein.
The invention also provides a substantially pure polypeptide characterized by having an amino acid sequence derived from carboxy terminal amino acids of a hedgehog protein and having at its amino terminus a G⇓CF cleavage site specifically recognized by a proteolytic activity of the carboxy terminal fragment of the native hedgehog polypeptide. This fragment is denoted the C-terminal fragment or polypeptide or “C”, herein. For example, in Drosophila this “C” polypeptide derives from the C-terminal domain of hedgehog precursor protein beginning at amino acid residue 258, wherein the full length C-terminal domain has a molecular weight of about 25 kD by non-reducing SDS-PAGE, a histidine residue at position 72, and has protease activity. The G⇓CF cleavage site specifically recognized by the proteolytic activity of the carboxy terminal fragment of the native hedgehog polypeptide is located at amino acid residues 257–259. As described above for the N fragment, now that the present invention has shown the precise cleavage recognition site for the autoproteolytic domain of hedgehog, those of skill in the art can readily discern the cleavage site in other hedgehog proteins thereby allowing the ready identification of any N or C polypeptide of any hedgehog precursor protein.
The “C” polypeptide of the invention is derived from the C-terminus of a hedgehog precursor protein, beginning at the autoproteolytic cleavage site identified at the GCF amino acid sequence, which in Drosophila corresponds to amino acids 257–259. In Drosophila the histidine residue found invariably at amino acid residue 329 of the native hedgehog protein, and at amino acid residue 72 of the C polypeptide, is essential for auto-proteolytic cleavage between amino acids 257 and 258 (G and C). Corresponding C-polypeptides of the invention will likewise contain a similarly located histidine residue which can be readily identified, such as by comparison to the Drosophila C-polypeptide. Among various species, the proteolytic domain can be characterized by the amino acid sequence-XTXXHLXX-(SEQ ID NO:36).
The C polypeptide of the invention, unlike N, does not significantly bind to heparin agarose. C is characterized by being released into the culture supernatant of cells expressing C polypeptide in vitro and by being localized diffusely in cells and embryos. Because C polypeptide diffuses freely, it would be detectable in various body fluids and tissues in a subject. Identification of C polypeptide expression near the midline of the neural tube, as described herein, provides a useful assay for neural tube closure in an embryo/fetus, for example. The presence of C polypeptide in amniotic fluid would be diagnostic of a disorder in which the neural tube may be malformed.
Altered levels of C polypeptide in cerebrospinal fluid may be indicative of neuro-degenerative disorders, for example. Because C polypeptide is released from the cell after synthesis and autoproteolysis of native hedgehog precursor polypeptide, tumors synthesizing and releasing high levels of C polypeptide would be detectable without prior knowledge of the exact location of the tumor.
C fragment is effective in inducing genes of the pituitary and anterior brain as well. In particular, induction is increased by the addition of a member of the TGF-β family of growth factors. For example, human activin in combination with C fragment may be effective in enhancing pituitary cell growth and activity or development. C fragment possesses cholesterol transferase activity thereby effecting precursor cleavage and transfer of a cholesterol moiety to N fragment, resulting in a biologically active N fragment.
C fragment is effective in inducing posterior markers of the brain by inhibiting N. Such a fragment is exemplified in Example 18 as ΔN-C. Therefore in another embodiment, the invention includes a polypeptide deleting amino acid residues 28–194 of X-bhh. (Autoproteolysis gives a C domain of 198–409 as well as a seven amino acid peptide, representing aa 24–27 and 195–197). This polypeptide blocks the activity of X-bhh and N in explants and reduces dorsoanterior structures in embryos. Also included are polynucleotide sequences encoding ΔN-C. ΔN-C is useful for increasing expression of posterior neural markers (e.g., En-2, Krox-20, Xlttbox-6) and decreasing expression of anterior neural markers (e.g., XANF-2, XAG-1, Otx-A) when desirable to do so to modulate neural patterning.
The term “substantially pure” as used herein refers to hedgehog N or C polypeptide which is substantially free of other proteins, lipids, carbohydrates, nucleic acids or other materials with which it is naturally associated. One skilled in the art can purify hedgehog N or C polypeptide using standard techniques for protein purification. The substantially pure polypeptide will yield a single major band on a non-reducing polyacrylamide gel. The purity of the hedgehog N or C polypeptide can also be determined by amino-terminal amino acid sequence analysis.
The invention includes a functional N or C polypeptide, and functional fragments thereof. As used herein, the term “functional polypeptide” or “functional fragment” refers to a polypeptide which possesses a biological function or activity which is identified through a defined functional assay and which is associated with a particular biologic, morphologic, or phenotypic alteration in the cell. Functional fragments of the hedgehog N or C polypeptide include fragments of N or C polypeptide as long as the activity, e.g., proteolytic activity or cholesterol transferase activity of C polypeptide remains. Smaller peptides containing the biological activity of N or C polypeptide are therefore included in the invention. The biological function, for example, can vary from a polypeptide fragment as small as an epitope to which an antibody molecule can bind to a large polypeptide which is capable of participating in the characteristic induction or programming of phenotypic changes within a cell. A “functional polynucleotide” denotes a polynucleotide which encodes a functional polypeptide as described herein.
Biologically active or functional fragments of hedgehog, as described herein, are included in the invention and can be identified as such by functional assays. For example, fragments of hedgehog are identified as inducing differentiation of neuronal cells; regulating differentiation of chondrocytes; able to complement a loss of function mutation of hedgehog, for example in a transgenic Drosophila; binding to Patched (Ptc); or having cholesterol transferase activity (e.g., C fragment). Fragments of the invention may be from about 30 to 450 amino acids in length; from about 50 to 300 amino acids in length; from about 75 to 250 amino acids in length; or from about 100 to 200 amino acids in length, as long as a biological activity of hedgehog is retained therein.
Minor modifications of the N or C polypeptide primary amino acid sequence may result in polypeptides which have substantially equivalent activity as compared to the N or C polypeptide described herein. Such modifications may be deliberate, as by site-directed mutagenesis, or may be spontaneous. All of the polypeptides produced by these modifications are included herein as long as the proteolytic activity of C polypeptide, for example, is present. Further, deletion of one or more amino acids can also result in a modification of the structure of the resultant molecule without significantly altering its activity. This can lead to the development of a smaller active molecule which would have broader utility. For example, it is possible to remove amino or carboxy terminal amino acids which may not be required for N or C polypeptide activity.
The N or C polypeptide of the invention also includes conservative variations of the polypeptide sequence. The term “conservative variation” as used herein denotes the replacement of an amino acid residue by another biologically similar residue. Examples of conservative variations include the substitution of one hydrophobic residue such as isoleucine, valine, leucine or methionine for another, or the substitution of one polar residue for another, such as the substitution of arginine for lysine, glutamic for aspartic acids, or glutamine for asparagine, and the like. The term “conservative variation” also includes the use of a substituted amino acid in place of an unsubstituted parent amino acid provided that antibodies raised to the substituted polypeptide also immunoreact with the unsubstituted polypeptide.
The N fragment of the invention includes both the active form of the polypeptide and the N fragment including the uncleaved signal sequence. For example, in Drosophila where the signal sequence is internal (at about amino acids 60–80), the entire uncleaved N fragment beginning at the initiating methionine is included in the invention. Those of skill in the art can readily ascertain the nature and location of the signal sequence by using, for example, the algorithm described in von Heijne, G., Nucl. Acids Res. 14:4683, (1986).
Hedgehog polypeptides of the invention include polypeptides having at least about 50%–100% homology with the hedgehog polypeptides provided herein, for example 52%, 64%, 68%, 70%, 75%, 80%, 85%, 90%, 95% and up to 100% homology. Preferably homologous polypeptides are derived from vertebrate species, most preferably mammalian species, such as humans.
The invention also provides an isolated polynucleotide sequence encoding a polypeptide having the amino acid sequence of N or C polypeptide of the invention. The term “isolated” as used herein includes polynucleotides substantially free of other nucleic acids, proteins, lipids, carbohydrates or other materials with which it is naturally associated. Polynucleotide sequences of the invention include DNA, cDNA and RNA sequences which encode N or C polypeptide. It is understood that all polynucleotides encoding all or a portion of N or C polypeptide are also included herein, as long as they encode a polypeptide with N or C polypeptide activity. Such polynucleotides include naturally occurring, synthetic, and intentionally manipulated polynucleotides. For example, N or C polypeptide polynucleotide may be subjected to site-directed mutagenesis. The polynucleotide sequence for N or C polypeptide also includes antisense sequences. The polynucleotides of the invention include sequences that are degenerate as a result of the genetic code. There are 20 natural amino acids, most of which are specified by more than one codon. Therefore, all degenerate nucleotide sequences are included in the invention as long as the amino acid sequence of N or C polypeptide encoded by the nucleotide sequence is functionally unchanged. In addition, the invention also includes a polynucleotide consisting essentially of a polynucleotide sequence encoding a polypeptide having an amino acid sequence of N or C and having at least one epitope for an antibody immunoreactive with N or C polypeptide.
The polynucleotide encoding N or C polypeptide includes the entire polypeptide or fragments thereof, as well as nucleic acid sequences complementary to that sequence. A complementary sequence may include an antisense nucleotide. When the sequence is RNA, the deoxynucleotides A, G, C, and T are replaced by ribonucleotides A, G, C, and U, respectively. Also included in the invention are fragments of the above-described nucleic acid sequences that are at least 15 bases in length, which is sufficient to permit the fragment to selectively hybridize to DNA that encodes the protein under physiological conditions.
Hedgehog encoding polynucleotides of the invention include nucleic acid sequences identified by hybridization to a hedgehog nucleic acid described herein. In nucleic acid hybridization reactions, the conditions used to achieve a particular level of stringency will vary, depending on the nature of the nucleic acids being hybridized. For example, the length, degree of complementarity, nucleotide sequence composition (e.g., GC v. AT content), and nucleic acid type (e.g., RNA v. DNA) of the hybridizing regions of the nucleic acids can be considered in selecting hybridization conditions. An additional consideration is whether one of the nucleic acids is immobilized, for example, on a filter.
An example of progressively higher stringency conditions is as follows: 2×SSC/0.1% SDS at about room temperature (hybridization conditions); 0.2×SSC/0.1% SDS at about room temperature (low stringency conditions); 0.2×SSC/0.1% SDS at about 42° C. (moderate stringency conditions); and 0.1×SSC at about 68° C. (high stringency conditions). Washing can be carried out using only one of these conditions, e.g., high stringency conditions, or each of the conditions can be used, e.g., for 10–15 minutes each, in the order listed above, repeating any or all of the steps listed. However, as mentioned above, optimal conditions will vary, depending on the particular hybridization reaction involved, and can be determined empirically.
DNA sequences of the invention can be obtained by several methods. For example, the DNA can be isolated using hybridization techniques which are well known in the art. These include, but are not limited to: 1) hybridization of genomic or cDNA libraries with probes to detect homologous nucleotide sequences; 2) antibody screening of expression libraries to detect cloned DNA fragments with shared structural features; and 3) PCR amplification of a desired nucleotide sequence using oligonucleotide primers.
Preferably the hedgehog, N, or C polynucleotide of the invention is derived from a vertebrate organism, and most preferably from human. Screening procedures which rely on nucleic acid hybridization make it possible to isolate any gene sequence from any organism, provided the appropriate probe is available. Oligonucleotide probes, which correspond to a part of the sequence encoding the protein in question, can be synthesized chemically. This requires that short, oligopeptide stretches of amino acid sequence must be known. The DNA sequence encoding the protein can be deduced from the genetic code, however, the degeneracy of the code must be taken into account. It is possible to perform a mixed addition reaction when the sequence is degenerate.
This includes a heterogeneous mixture of denatured double-stranded DNA. For such screening, hybridization is preferably performed on either single-stranded DNA or denatured double-stranded DNA. Hybridization is particularly useful in the detection of cDNA clones derived from sources where an extremely low amount of mRNA sequences relating to the polypeptide of interest are present. In other words, by using stringent hybridization conditions directed to avoid non-specific binding, it is possible, for example, to allow the autoradiographic visualization of a specific cDNA clone by the hybridization of the target DNA to that single probe in the mixture which is its complete complement (Wallace, et al., Nucl. Acid Res., 9:879, 1981).
The development of specific DNA sequences encoding hedgehog can also be obtained by: 1) isolation of double-stranded DNA sequences from the genomic DNA; 2) chemical manufacture of a DNA sequence to provide the necessary codons for the polypeptide of interest; and 3) in vitro synthesis of a double-stranded DNA sequence by reverse transcription of mRNA isolated from a eukaryotic donor cell. In the latter case, a double-stranded DNA complement of mRNA is eventually formed which is generally referred to as cDNA.
Of the three above-noted methods for developing specific DNA sequences for use in recombinant procedures, the isolation of genomic DNA isolates is the least common. This is especially true when it is desirable to obtain the microbial expression of mammalian polypeptides due to the presence of introns.
The synthesis of DNA sequences is frequently the method of choice when the entire sequence of amino acid residues of the desired polypeptide product is known. When the entire sequence of amino acid residues of the desired polypeptide is not known, the direct synthesis of DNA sequences is not possible and the method of choice is the synthesis of cDNA sequences. Among the standard procedures for isolating cDNA sequences of interest is the formation of plasmid- or phage-carrying cDNA libraries which are derived from reverse transcription of mRNA which is abundant in donor cells that have a high level of genetic expression. When used in combination with polymerase chain reaction technology, even rare expression products can be cloned. In those cases where significant portions of the amino acid sequence of the polypeptide are known, the production of labeled single or double-stranded DNA or RNA probe sequences duplicating a sequence putatively present in the target cDNA may be employed in DNA/DNA hybridization procedures which are carried out on cloned copies of the cDNA which have been denatured into a single-stranded form (Jay, et al., Nucl. Acid Res., 11:2325, 1983).
A preferred method for obtaining genomic DNA, for example, is Polymerase Chain Reaction (PCR), which relies on an in vitro method of nucleic acid synthesis by which a particular segment of DNA is specifically replicated. Two oligonucleotide primers that flank the DNA fragment to be amplified are utilized in repeated cycles of heat denaturation of the DNA, annealing of the primers to their complementary sequences, and extension of the annealed primers with DNA polymerase. These primers hybridize to opposite strands of the target sequence and are oriented so that DNA synthesis by the polymerase proceeds across the region between the primers. Since the extension products themselves are also complementary to and capable of binding primers, successive cycles of amplification essentially double the amount of the target DNA synthesized in the previous cycle. The result is an exponential accumulation of the specific target fragment, approximately 2n, where n is the number of cycles of amplification performed (see PCR Protocols, Eds. Innis, et al., Academic Press, Inc., 1990, incorporated herein by reference).
A cDNA expression library, such as λgt11, can be screened indirectly for hedgehog, N, or C polypeptides having at least one epitope, using antibodies specific for hedgehog, N, or C. Such antibodies can be either polyclonally or monoclonally derived and used to detect expression product indicative of the presence of the desired hedgehog cDNA.
The polynucleotide sequence for hedgehog, N, or C, also includes sequences complementary to the polynucleotide encoding hedgehog, N or C (antisense sequences). Antisense nucleic acids are DNA or RNA molecules that are complementary to at least a portion of a specific mRNA molecule (Weintraub, Scientific American, 262:40, 1990). The invention embraces all antisense polynucleotides capable of inhibiting production of hedgehog, N, or C polypeptide. In the cell, the antisense nucleic acids hybridize to the corresponding mRNA, forming a double-stranded molecule. The antisense nucleic acids interfere with the translation of the mRNA since the cell will not translate a mRNA that is double-stranded. Antisense oligomers of about 15 nucleotides are preferred, since they are easily synthesized and are less likely to cause problems than larger molecules when introduced into the target hedgehog, N, or C-producing cell. The use of antisense methods to inhibit the translation of genes is well known in the art (Marcus-Sakura, Anal. Biochem. 172:289, 1988). Inhibition of target nucleotide would be desirable, for example, in inhibiting cell-proliferative disorders, such as certain tumors, which are mediated by hedgehog, N or C.
In addition, ribozyme nucleotide sequences for hedgehog, N or C are included in the invention. Ribozymes are RNA molecules possessing the ability to specifically cleave other single-stranded RNA in a manner analogous to DNA restriction endonucleases. Through the modification of nucleotide sequences which encode these RNAs, it is possible to engineer molecules that recognize specific nucleotide sequences in an RNA molecule and cleave it (Cech, J. Amer. Med. Assn., 260:3030, 1988). A major advantage of this approach is that, because they are sequence-specific, only mRNAs with particular sequences are inactivated.
There are two basic types of ribozymes namely, tetrahymena-type (Hasselhoff, Nature, 334:585, 1988) and “hammerhead”-type. Tetrahymena-type ribozymes recognize sequences which are four bases in length, while “hammerhead”-type ribozymes recognize base sequences 11–18 bases in length. The longer the recognition sequence, the greater the likelihood that sequence will occur exclusively in the target mRNA species. Consequently, hammerhead-type ribozymes are preferable to tetrahymena-type ribozymes for inactivating a specific mRNA species and 18-based recognition sequences are preferable to shorter recognition sequences.
DNA sequences encoding hedgehog, N or C can be expressed in vitro by DNA transfer into a suitable host cell. “Host cells” are cells in which a vector can be propagated and its DNA expressed. The term also includes any progeny of the subject host cell. It is understood that all progeny may not be identical to the parental cell since there may be mutations that occur during replication. However, such progeny are included when the term “host cell” is used. Methods of stable transfer, meaning that the foreign DNA is continuously maintained in the host, are known in the art.
In the present invention, the hedgehog, N or C polynucleotide sequences maybe inserted into a recombinant expression vector. The term “recombinant expression vector” refers to a plasmid, virus or other vehicle known in the art that has been manipulated by insertion or incorporation of the hedgehog, N or C genetic sequences. Such expression vectors contain a promoter sequence which facilitates the efficient transcription of the inserted genetic sequence of the host. The expression vector typically contains an origin of replication, a promoter, as well as specific genes which allow phenotypic selection of the transformed cells. Vectors suitable for use in the present invention include, but are not limited to the T7-based expression vector for expression in bacteria (Rosenberg, et al., Gene, 56:125, 1987), the pMSXND expression vector for expression in mammalian cells (Lee and Nathans, J. Biol. Chem., 263:3521, 1988) and baculovirus-derived vectors for expression in insect cells. The DNA segment can be present in the vector operably linked to regulatory elements, for example, a promoter (e.g., T7, metallothionein I, or polyhedrin promoters).
Polynucleotide sequences encoding hedgehog, N or C can be expressed in either prokaryotes or eukaryotes, although post-translational modification of eukaryotically derived polypeptides, such as carboxylation, would occur in a eukaryotic host. Hosts can include microbial, yeast, insect and mammalian organisms. Methods of expressing DNA sequences having eukaryotic or viral sequences in prokaryotes are well known in the art. Biologically functional viral and plasmid DNA vectors capable of expression and replication in a host are known in the art. Such vectors are used to incorporate DNA sequences of the invention.
Methods which are well known to those skilled in the art can be used to construct expression vectors containing the hedgehog, N or C coding sequence and appropriate transcriptional/translational control signals. These methods include in vitro recombinant DNA techniques, synthetic techniques, and in vivo recombination/genetic techniques. See, for example, the techniques described in Maniatis, et al., 1989 Molecular Cloning A Laboratory Manual, Cold Spring Harbor Laboratory, N.Y.
A variety of host-expression vector systems may be utilized to express the hedgehog, N or C coding sequence. These include but are not limited to microorganisms such as bacteria transformed with recombinant bacteriophage DNA, plasmid DNA or cosmid DNA expression vectors containing the hedgehog, N or C coding sequence; yeast transformed with recombinant yeast expression vectors containing the hedgehog, N or C coding sequence; plant cell systems infected with recombinant virus expression vectors (e.g., cauliflower mosaic virus, CaMV; tobacco mosaic virus, TMV) or transformed with recombinant plasmid expression vectors (e.g., Ti plasmid) containing the Hedgehog, N or C coding sequence; insect cell systems infected with recombinant virus expression vectors (e.g., baculovirus) containing the hedgehog, N or C coding sequence; or animal cell systems infected with recombinant virus expression vectors (e.g., retroviruses, adenovirus, vaccinia virus) containing the hedgehog, N or C coding sequence, or transformed animal cell systems engineered for stable expression.
Depending on the host/vector system utilized, any of a number of suitable transcription and translation elements, including constitutive and inducible promoters, transcription enhancer elements, transcription terminators, etc., may be used in the expression vector (see e.g., Bitter, et al., 1987, Methods in Enzymology, 153:516–544). For example, when cloning in bacterial systems, inducible promoters such as pL of bacteriophage γ, plac, ptrp, ptac (ptrp-lac hybrid promoter) and the like may be used. When cloning in mammalian cell systems, promoters derived from the genome of mammalian cells (e.g., metallothionein promoter) or from mammalian viruses (e.g., the retrovirus long terminal repeat; the adenovirus late promoter; the vaccinia virus 7.5K promoter) may be used. Promoters produced by recombinant DNA or synthetic techniques may also be used to provide for transcription of the inserted hedgehog, N or C coding sequence.
In bacterial systems a number of expression vectors may be advantageously selected depending upon the use intended for the expressed. For example, when large quantities of hedgehog, N or C are to be produced, vectors which direct the expression of high levels of fusion protein products that are readily purified may be desirable. Those which are engineered to contain a cleavage site to aid in recovering are preferred. Such vectors include but are not limited to the E. coli expression vector pUR278 (Ruther, et al., EMBO J., 2:1791, 1983), in which the Hedgehog, N or C coding sequence may be ligated into the vector in frame with the lac Z coding region so that a hybrid -lac Z protein is produced; pIN vectors (Inouye and Inouye, Nucleic Acids Res., 13:3101, 1985; Van Heeke and Schuster, J. Biol. Chem. 264:5503; 1989) and the like.
In yeast, a number of vectors containing constitutive or inducible promoters may be used. For a review see, Current Protocols in Molecular Biology, Vol. 2, 1988, Ed. Ausubel, et al., Greene Publish. Assoc. & Wiley Interscience, Ch. 13; Grant, et al., 1987, Expression and Secretion Vectors for Yeast, in Methods in Enzymology, Eds. Wu and Grossman, 31987, Acad. Press, N.Y., Vol. 153, pp. 516–544; Glover, 1986, DNA Cloning, Vol. II, IRL Press, Wash., D.C., Ch. 3; and Bitter, 1987, Heterologous Gene Expression in Yeast, Methods in Enzymology, Eds. Berger and Kimmel, Acad. Press, N.Y., Vol. 152, pp. 673–684; and The Molecular Biology of the Yeast Saccharomyces, 1982, Eds. Strathern et al., Cold Spring Harbor Press, Vols. I and II. A constitutive yeast promoter such as ADH or LEU2 or an inducible promoter such as GAL may be used (Cloning in Yeast, Ch. 3, R. Rothstein In: DNA Cloning Vol. 11, A Practical Approach, Ed. D M Glover, 1986, IRL Press, Wash., D.C.). Alternatively, vectors may be used which promote integration of foreign DNA sequences into the yeast chromosome.
In cases where plant expression vectors are used, the expression of the hedgehog, N or C coding sequence may be driven by any of a number of promoters. For example, viral promoters such as the 35S RNA and 19S RNA promoters of CaMV (Brisson, et al., Nature, 310:511, 1984), or the coat protein promoter to TMV (Takamatsu, et al., EMBO J., 6:307, 1987) may be used; alternatively, plant promoters such as the small subunit of RUBISCO (Coruzzi, et al., EMBO J., 3:1671–1680, 1984; Broglie, et al., Science, 224:838, 1984); or heat shock promoters, e.g., soybean hsp17.5-E or hsp17.3-B (Gurley, et al., Mol. Cell. Biol., 6:559, 1986) may be used. These constructs can be introduced into plant cells using Ti plasmids, Ri plasmids, plant virus vectors, direct DNA transformation, microinjection, electroporation, etc. For reviews of such techniques see, for example, Weissbach and Weissbach, 1988, Methods for Plant Molecular Biology, Academic Press, NY, Section VIII, pp. 421–463; and Grierson and Corey, 1988, Plant Molecular Biology, 2d Ed., Blackie, London, Ch. 7–9.
An alternative expression system which could be used to express is an insect system. In one such system, Autographa californica nuclear polyhedrosis virus (AcNPV) is used as a vector to express foreign genes. The virus grows in Spodoptera frugiperda cells. The hedgehog, N or C coding sequence may be cloned into non-essential regions (for example the polyhedrin gene) of the virus and placed under control of an AcNPV promoter (for example the polyhedrin promoter). Successful insertion of the hedgehog, N or C coding sequence will result in inactivation of the polyhedrin gene and production of non-occluded recombinant virus (i.e., virus lacking the proteinaceous coat coded for by the polyhedrin gene). These recombinant viruses are then used to infect Spodoptera frugiperda cells in which the inserted gene is expressed. (e.g., see Smith, et al., J. Viol., 46:584, 1983; Smith, U.S. Pat. No. 4,215,051).
Eukaryotic systems, and preferably mammalian expression systems, allow for proper post-translational modifications of expressed mammalian proteins to occur. Eukaryotic cells which possess the cellular machinery for proper processing of the primary transcript, glycosylation, phosphorylation, and advantageously, secretion of the gene product may be used as host cells for the expression of hedgehog, N or C. Mammalian cell lines may be preferable. Such host cell lines may include but are not limited to CHO, VERO, BHK, HeLa, COS, MDCK, −293, and WI38.
Mammalian cell systems which utilize recombinant viruses or viral elements to direct expression may be engineered. For example, when using adenovirus expression vectors, the hedgehog, N or C coding sequence may be ligated to an adenovirus transcription/translation control complex, e.g., the late promoter and tripartite leader sequence. This chimeric gene may then be inserted in the adenovirus genome by in vitro or in vivo recombination. Insertion in a non-essential region of the viral genome (e.g., region E1 or E3) will result in a recombinant virus that is viable and capable of expressing the protein in infected hosts (e.g., see Logan and Shenk, Proc. Natl. Acad. Sci. USA, 81:3655, 1984). Alternatively, the vaccinia virus 7.5K promoter may be used. (e.g., see, Mackett, et al., Proc. Natl. Acad. Sci. USA, 79:7415, 1982; Mackett, et al., J. Virol, 49: 857, 1984; Panicali, et al., Proc. Natl. Acad. Sci. USA, 79:4927, 1982). Of particular interest are vectors based on bovine papilloma virus which have the ability to replicate as extrachromosomal elements (Sarver, et al., Mol. Cell. Biol., 1:486, 1981). Shortly after entry of this DNA into mouse cells, the plasmid replicates to about 100 to 200 copies per cell. Transcription of the inserted cDNA does not require integration of the plasmid into the host's chromosome, thereby yielding a high level of expression. These vectors can be used for stable expression by including a selectable marker in the plasmid, such as, for example, the neo gene. Alternatively, the retroviral genome can be modified for use as a vector capable of introducing and directing the expression of the hedgehog, N or C gene in host cells (Cone and Mulligan, Proc. Natl. Acad. Sci. USA, 81:6349, 1984). High level expression may also be achieved using inducible promoters, including, but not limited to, the metallothionine IIA promoter and heat shock promoters.
For long-term, high-yield production of recombinant proteins, stable expression is preferred. Rather than using expression vectors which contain viral origins of replication, host cells can be transformed with the hedgehog, N or C cDNA controlled by appropriate expression control elements (e.g., promoter, enhancer, sequences, transcription terminators, polyadenylation sites, etc.), and a selectable marker. The selectable marker in the recombinant plasmid confers resistance to the selection and allows cells to stably integrate the plasmid into their chromosomes and grow to form foci which in turn can be cloned and expanded into cell lines. For example, following the introduction of foreign DNA, engineered cells may be allowed to grow for 1–2 days in an enriched media, and then are switched to a selective media. A number of selection systems may be used, including but not limited to the herpes simplex virus thymidine kinase (Wigler, et al., Cell, 11: 223, 1977), hypoxanthine-guanine phosphoribosyltransferase (Szybalska and Szybalski, Proc. Natl. Acad. Sci. USA, 48:2026, 1962), and adenine phosphoribosyltransferase (Lowy, et al., Cell, 22: 817, 1980) genes can be employed in tk, hgprt or aprt cells respectively. Also, antimetabolite resistance can be used as the basis of selection for dhfr, which confers resistance to methotrexate (Wigler, et al., Natl. Acad. Sci. USA4, 77: 3567, 1980; O'Hare, et al., Proc. Natl. Acad. Sci. USA, 78: 1527, 1981); gpt, which confers resistance to mycophenolic acid (Mulligan and Berg, Proc. Natl. Acad. Sci. USA, 78: 2072, 1981; neo, which confers resistance to the aminoglycoside G-418 (Colberre-Garapin, et al., J. Mol. Biol., 150:1, 1981); and hygro, which confers resistance to hygromycin (Santerre, et al., Gene, 30:147, 1984) genes. Recently, additional selectable genes have been described, namely trpB, which allows cells to utilize indole in place of tryptophan; hisD, which allows cells to utilize histinol in place of histidine (Hartman and Mulligan, Proc. Natl. Acad. Sci. USA, 85:8047, 1988); and ODC (ornithine decarboxylase) which confers resistance to the ornithine decarboxylase inhibitor, 2-(difluoromethyl)-DL-ornithine, DFMO (McConlogue L., 1987, In: Current Communications in Molecular Biology, Cold Spring Harbor Laboratory ed.).
Transformation of a host cell with recombinant DNA may be carried out by conventional techniques as are well known to those skilled in the art. Where the host is prokaryotic, such as E. coli, competent cells which are capable of DNA uptake can be prepared from cells harvested after exponential growth phase and subsequently treated by the CaCl2 method using procedures well known in the art. Alternatively, MgCl2 or RbCl can be used. Transformation can also be performed after forming a protoplast of the host cell if desired.
When the host is a eukaryote, such methods of transfection of DNA as calcium phosphate co-precipitates, conventional mechanical procedures such as microinjection, electroporation, insertion of a plasmid encased in liposomes, or virus vectors may be used. Eukaryotic cells can also be cotransformed with DNA sequences encoding the hedgehog, N or C of the invention, and a second foreign DNA molecule encoding a selectable phenotype, such as the herpes simplex thymidine kinase gene. Another method is to use a eukaryotic viral vector, such as simian virus 40 (SV40) or bovine papilloma virus, to transiently infect or transform eukaryotic cells and express the protein. (see for example, Eukaryotic Viral Vectors, Cold Spring Harbor Laboratory, Gluzman ed., 1982).
Isolation and purification of microbial expressed polypeptide, or fragments thereof, provided by the invention, may be carried out by conventional means including preparative chromatography and immunological separations involving monoclonal or polyclonal antibodies.
The invention includes antibodies immunoreactive with or which bind to hedgehog, N or C polypeptide or functional fragments thereof. Antibody which consists essentially of pooled monoclonal antibodies with different epitopic specificities, as well as distinct monoclonal antibody preparations are provided. Monoclonal antibodies are made from antigen containing fragments of the protein by methods well known to those skilled in the art (Kohler, et al., Nature, 256:495, 1975). The term antibody as used in this invention is meant to include intact molecules as well as fragments thereof, such as Fab and F(ab′)2, which are capable of binding an epitopic determinant on hedgehog, N or C. The antibodies of the invention include antibodies which bind to the N or C polypeptide and which bind with immunoreactive fragments N or C.
The term “antibody” as used in this invention includes intact molecules as well as fragments thereof, such as Fab, F(ab′)2, and Fv which are capable of binding the epitopic determinant. These antibody fragments retain some ability to selectively bind with its antigen or receptor and are defined as follows:
Methods of making these fragments are known in the art. (See for example, Harlow and Lane, Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory, New York (1988), incorporated herein by reference).
As used in this invention, the term “epitope” means any antigenic determinant on an antigen to which the paratope of an antibody binds. Epitopic determinants usually consist of chemically active surface groupings of molecules such as amino acids or sugar side chains and usually have specific three dimensional structural characteristics, as well as specific charge characteristics.
Antibodies which bind to the hedgehog, N or C polypeptide of the invention can be prepared using an intact polypeptide or fragments containing small peptides of interest as the immunizing antigen. The polypeptide such as N or C, or fragments thereof used to immunize an animal can be derived from translated cDNA or chemical synthesis which can be conjugated to a carrier protein, if desired.
Such commonly used carriers which are chemically coupled to the peptide include keyhole limpet hemocyanin (KLH), thyroglobulin, bovine serum albumin (BSA), and tetanus toxoid. The coupled peptide is then used to immunize the animal (e.g., a mouse, a rat, or a rabbit). If desired, polyclonal or monoclonal antibodies can be further purified, for example, by binding to and elution from a matrix to which the polypeptide or a peptide to which the antibodies were raised is bound. Those of skill in the art will know of various techniques common in the immunology arts for purification and/or concentration of polyclonal antibodies, as well as monoclonal antibodies (See for example, Coligan, et al., Unit 9, Current Protocols in Immunology, Wiley Interscience, 1991, incorporated by reference).
It is also possible to use the anti-idiotype technology to produce monoclonal antibodies which mimic an epitope. For example, an anti-idiotypic monoclonal antibody made to a first monoclonal antibody will have a binding domain in the hypervariable region which is the “image” of the epitope bound by the first monoclonal antibody.
Antibodies as described herein as having specificity for N polypeptide, e.g., Ab1 (residues 83–160), are useful for specific identification of cells or tissues expressing the N fragment of hedgehog. Similarly, antibodies described herein as having specificity for C polypeptide, e.g., Ab2 (residues 300–391), are useful for specific identification of cells or tissues expressing the C fragment of hedgehog. Both antibodies, naturally, will also detect native hedgehog polypeptide.
The N and C-specific antibodies of the invention are useful for purification of N and C polypeptide, respectively, especially using the antibodies immobilized on solid phase. By contacting a sample with anti-N antibody, both N and native hedgehog polypeptides can be isolated. By next contacting the sample removed by anti-N antibodies, with anti-C antibodies, the native hedgehog polypeptide is removed, thus allowing purification of N polypeptide. In a similar manner, C polypeptide can be antibody purified from a sample.
Monoclonal antibodies of the invention are suited for use, for example, in immunoassays in which they can be utilized in liquid phase or bound to a solid phase carrier. In addition, the monoclonal antibodies in these immunoassays can be detectably labeled in various ways. Examples of types of immunoassays which can utilize monoclonal antibodies of the invention are competitive and non-competitive immunoassays in either a direct or indirect format. Examples of such immunoassays are the radioimmunoassay (RIA) and the sandwich (immunometric) assay. Detection of the antigens using the monoclonal antibodies of the invention can be done utilizing immunoassays which are run in either the forward, reverse, or simultaneous modes, including immunohistochemical assays on physiological samples. Those of skill in the art will know, or can readily discern, other immunoassay formats without undue experimentation.
The term “immunometric assay” or “sandwich immunoassay”, includes simultaneous sandwich, forward sandwich and reverse sandwich immunoassays. These terms are well understood by those skilled in the art. Those of skill will also appreciate that antibodies according to the present invention will be useful in other variations and forms of assays which are presently known or which may be developed in the future. These are intended to be included within the scope of the present invention.
Monoclonal antibodies can be bound to many different carriers and used to detect the presence of N or C polypeptide. Examples of well-known carriers include glass, polystyrene, polypropylene, polyethylene, dextran, nylon, amylases, natural and modified celluloses, polyacrylamides, agaroses and magnetite. The nature of the carrier can be either soluble or insoluble for purposes of the invention. Those skilled in the art will know of other suitable carriers for binding monoclonal antibodies, or will be able to ascertain such using routine experimentation.
For purposes of the invention, N or C polypeptide may be detected by the monoclonal antibodies when present in biological fluids and tissues. Any sample containing a detectable amount of N or C can be used. A sample can be a liquid such as urine, saliva, cerebrospinal fluid, blood, serum and the like, or a solid or semi-solid such as tissues, feces, and the like, or, alternatively, a solid tissue such as those commonly used in histological diagnosis. C polypeptide in particular is detectable in biological samples, since it tends to diffuse more readily than N polypeptide.
In performing the assays it may be desirable to include certain “blockers” in the incubation medium (usually added with the labeled soluble antibody). The “blockers” are added to assure that non-specific proteins, proteases, or anti-heterophilic immunoglobulins to anti-C or N immunoglobulins present in the experimental sample do not cross-link or destroy the antibodies on the solid phase support, or the radiolabeled indicator antibody, to yield false positive or false negative results. The selection of “blockers” therefore may add substantially to the specificity of the assays described in the present invention.
The invention also provides a method for modulating proliferation or differentiation of neuronal cells comprising contacting the cells with a hedgehog polypeptide. The hedgehog polypeptide may be a native hedgehog polypeptide, or a N or C polypeptide, or functional fragments thereof. Preferably, the modulation is induction of proliferation or differentiation of a particular cell type. This can involve either synergistic positive induction of neuronal cells by N, or negative modulation by delta N-C for example (Lai, et al., Development 121:2349, 1995). Delta N-C enhances expression of posterier relative to anterior neural genes and does so through inhibition of N (see EXAMPLE 18 and
Previous studies with the rat hedgehog gene showed that co-culture of cells expressing rat hedgehog precursor gene, with explant from neural tube, was sufficient to induce formation of motor neurons and floor plate from the explant (Jessesl, T., and Dodd, J., In Cell-Cell Signaling in Vertebrate Development (ed. E. J. Robertson, et al., pp 139–155, San Diego, Calif.), 1993). Therefore, based on the Examples herein showing that hedgehog is expressed near the Doorplate of the ventral midline of the neural tube and notochord, neuronal cells substantially derived from floor plate neuronal cells can be induced by contacting the cells with hedgehog, N or C polypeptide. As used herein, the term “substantially derived”, refers to those cells from the floor plate or proximate to the floor plate. For example, such cells include motor neurons and dopaminergic neurons. Those of skill in the art will be able to identify other neuronal cells substantially derived from the floor plate. Preferably the cells are vertebrate cells and most preferably, human cells.
In addition, as described herein in the Examples, hedgehog, and particularly C fragment, induces the expression of pituitary genes. Hedgehog is also effective in inducing anterior brain gene expression as exemplified by the OTX-A marker. Further, the addition of a TGF-β family member, for example activin, may be used to further induce expression of such genes. Other TGF-β family members will be known to those of skill in the art. This apparent synergy of hh fragments with TGF-β family members occurs through the TGF-β protein inducing expression of neural inducers such as noggin and follistatin. The hh fragment then synergizes with these inducers to pattern neural gene expression.
hh fragments may also be useful as nerve-sparing agents or in restoring or promoting appropriate patterning during the healing of major limb trauma. In addition, the N and C fragments may be useful in the area of genetic counseling. Specifically, familial midline defects such as cyclopia, polydactyly or neural tube defects may be diagnosed by mapping close to hh. Since autoproteolytic defects may be responsible for the disorders, N or C therapy could be provided.
The invention also provides an autoproteolytic fusion protein comprising a first polypeptide including the proteolytic domain of the C polypeptide of the invention, a cleavage site recognized by the first polypeptide, and a second polypeptide. (It is understood that the first and second polypeptides can be reversed.) The auto-proteolytic activity of the native hedgehog protein is found entirely within the C polypeptide, therefore, the C polypeptide is useful for producing a fusion polypeptide which can then be cleaved at the junction of the C polypeptide and the second polypeptide. The fusion protein may optionally have a purification tag, such as a poly-histidine tag for isolation on a nickel column, or an antibody epitope tag, preferably on the C fragment. The cleavage site includes the sequence “GCF”, which is recognized by the proteolytic domain of the C polypeptide and is utilized to cleave the second polypeptide from the C fragment. Also included in the invention is a polynucleotide encoding the fusion protein of the invention.
The invention also provides a method for producing an autoproteolytic fusion protein comprising operably linking a first polynucleotide, wherein the first polynucleotide encodes a first polypeptide including the proteolytic domain of the C polypeptide of the invention and the cleavage site recognized by the proteolytic domain, and a second polynucleotide encoding a second polypeptide. As described above, the fusion protein may also include a carrier peptide and/or a purification tag.
The C polypeptide or functional fragment thereof is useful as a fusion partner to cause lipophilic modification and tethering of other proteins in vivo or in vitro. Such fusion proteins may be desirable for factors whose activity is required in a localized manner, either by targeting DNA constructs to specific cells or by introducing cells transfected with specific DNA constructs, for example. It may be desirable to lipid-modify a normally secreted protein in order to produce a cell-associated protein. For example, it may be desirable to produce a viral antigen that remains cell associated. Specifically, cholesterol is covalently attached to the N-terminal protein during autoprocessing and the C polypeptide acts as an intramolecular cholesterol transferase.
Alternatively, the C polypeptide or functional fragments thereof can be used as a fusion partner with a protein of interest (e.g., Protein X fused to hh-C domain). Such fusions form thioesters at the junction between Protein X and hh-C (via an S to N shift). The thioesters are then available as substrates for a peptide ligation reaction in which any peptide or protein having an amino terminal cysteine (Peptide Y) is added and undergoes spontaneous rearrangement (S to N shift) that generates a stable peptide bond between Protein X and Peptide Y (Protein X-peptide bond-Peptide Y). For example, a protein that is toxic when produced in vivo could be produced in vitro using the hh-C domain fusion protein method.
The fusion polypeptide may also include an optional carrier peptide. The “carrier peptide”, or signal sequence, is located at the amino terminal end of the fusion peptide sequence. In the case of eukaryotes, the carrier peptide is believed to function to transport the fusion polypeptide across the endoplasmic reticulum. The secretory protein is then transported through the Golgi apparatus, into secretory vesicles and into the extracellular space or, preferably, the external environment. Carrier peptides which can be utilized according to the invention include pre-pro peptides which contain a proteolytic enzyme recognition site. Acceptable carrier peptides include the amino terminal pro-region of calcitonin or other hormones, which undergo cleavage at the flanking dibasic sites. However, it should be noted that the invention is not limited to the use of any particular peptide as a carrier. Other carrier peptides are known to those skilled in the art or can be readily ascertained without undue experimentation.
In one embodiment of the invention, a carrier peptide which is a signal sequence is included in the expression vector, specifically located adjacent to the N-terminal end of the fusion polypeptide. This signal sequence allows the fusion protein to be directed toward the endoplasmic reticulum. Typically, the signal sequence consists of a leader of from about 16 to about 29 amino acids, starting with two or three polar residues and continuing with a high content of hydrophobic amino acids; there is otherwise no detectable conservation of sequence known. Such signal sequences are known to those of skill in the art, and include the naturally occurring signal sequence derived from a hedgehog protein.
The fusion polypeptide of the invention includes a polypeptide encoded by a structural gene, preferably at the amino-terminus of the fusion polypeptide. Any structural gene is expressed in conjunction with the C-polypeptide (polynucleotide) and optionally a carrier peptide. The structural gene is operably linked with the carrier in an expression vector so that the fusion polypeptide is expressed as a single unit.
The identification of the autoproteolysis of hedgehog into the N and C domains is useful in a screening method to identify compounds or compositions which affect this processing activity. Thus, in another embodiment, the invention provides a method for identifying a composition which affects hh processing, which can be determined by activity or gene expression, comprising incubating the components, which include the composition to be tested (e.g., a drug, a small molecule, a protein) and a hh polypeptide or a recombinant cell expressing hedgehog or a gene encoding a C domain or functional fragment thereof operably linked to an N domain or functional fragment thereof, under conditions sufficient to allow the components to interact, then subsequently measuring the effect the composition has on hedgehog activity or expression. Fragments of hedgehog polypeptide or polynucleotide can be used in the method of the invention as long as autoproteolytic activity remains (e.g., the construct exemplified in
The identification of the lipid modification of the N domain of hedgehog by the C domain, resulting in a biologically active N domain, is useful, in a screening method to identify compounds or compositions which affect the cholesterol transferase/processing activity of hedgehog. In a broader aspect, the modification may be a general sterol or lipid modification, and not limited to cholesterol. Thus, in another embodiment, the invention provides a method for identifying a composition which affects hh biological activity, which can be determined by activity or lipid modification (e.g., cholesterol), comprising incubating the components, which include the composition to be tested (e.g., a drug, a small molecule, a protein) and a hh polypeptide or a recombinant cell expressing hedgehog or a gene encoding a C domain or functional fragment thereof operably linked to an N domain or functional fragment thereof, under conditions sufficient to allow the components to interact, then subsequently measuring the effect the composition has on hedgehog activity. Fragments of hedgehog polypeptide or polynucleotide can be used in the method of the invention as long as cholesterol transferase activity remains, for example. The effect on hh may be either inhibitory or stimulatory. For example, one can determine whether the N domain is associated with the cell, or whether the N domain is secreted into the medium, in other words, whether incomplete processing and modification has occurred. Such methods for determining the effect of the compound or composition on hh processing include those described herein (see Example 10,
As used herein, “hh activity” as described in the screening method refers preferably to auto-proteolytic activity. However, it is understood, that one of skill in the art could use the above-described screening assay to identify a composition having an affect on other hh activities, for example, zinc hydrolase activity or cholesterol transferase activity; or induction or regulation of differentiation of neuronal cells or chondrocytes. Appropriate assays for determining the effect on such activities will be known to those of skill in the art. Example 19 provides lipophilic modification assays useful in the described screening methods above.
Now that the present invention describes the cholesterol modification of N by C, it is possible to design various diagnostic and therapeutic approaches for treatment of hh associated disorders due to defective or altered sterol modification. For example, Smith-Lemli-Optiz syndrome (SLOS) is characterized by a loss of hh function and a sterol profile indicating a cholesterol deficiency. Therefore, SLOS may be diagnosed and/or treated based on the cholesterol profile. Further, a defect in Desert hh in the testes is associated with male sterility (M. Bitgood, L. Shen, A. P. McMahon, Current Biology 6, 298, 1996; A. Vortkamp et al., Science 273, 613, 1996), consequently, it may be possible to design male contraceptives based on defective cholesterol modification of hh. On the other hand, if sterility or decreased fertility was desirable, hh cholesterol transferase activity could be altered to reduce cholesterol modification. Processing of the C and N fragments of hh is required for hh activity, therefore alterations in cholesterol modification of the amino terminal fragment may also be related to developmental defects in vertebrate embryos.
Another aspect of the present invention concerns three-dimensional molecular models of the subject hedgehog proteins, and their use as templates for the design of agents able to inhibit or potentiate at least one biological activity of the hedgehog, particularly the autoproteolytic. An integral step to our approach to designing inhibitors of the subject hedgehog proteins, for example, involves construction of computer graphics models of the hedgehog protein which can be used to design pharmacophores by rational drug design. For instance, for an inhibitor to interact optimally with the subject proteolytic domain of hedgehog, it will generally be desirable that it have a shape which is at least partly complimentary to that of a particular binding site of the enzyme, as for example those portions of the human hedgehog protein which are involved in the autoproteolytic activity. Additionally, other factors, including electrostatic interactions, hydrogen bonding, hydrophobic interactions, desolvation effects, and cooperative motions of ligand and enzyme, all influence the binding effect and should be taken into account in attempts to design bioactive inhibitors.
A computer-generated molecular model of the subject hedgehog proteins can be created. In preferred embodiments, at least the Cα-carbon positions of the hedgehog sequence of interest are mapped to a particular coordinate pattern, such as the coordinates for hedgehog determined by x-ray crystallography, by homology modeling, and the structure of the protein and velocities of each atom are calculated at a simulation temperature (To) at which the docking simulation is to be determined. Typically, such a protocol involves primarily the prediction of side-chain conformations in the modeled protein, while assuming a main-chain trace taken from a tertiary structure such as provided in x-crystallographic model described herein. Computer programs for performing energy minimization routines are commonly used to generate molecular models. For example, both the CHARMM (Brooks et al. (1983) J Comput Chem 4:187–217) and AMBER (Weiner et al (1981) J. Comput. Chem. 106: 765) algorithms handle all of the molecular system setup, force field calculation, and analysis (see also, Eisenfield et al. (1991) Am J Physiol 261:C376–386; Lybrand (1991) J Pharm Belg 46:49–54; Froimowitz (1990) Biotechniques 8:640–644; Burbam et al. (1990) Proteins 7:99–111; Pedersen (1985) Environ Health Perspect 61:185–190; and Kini et al. (1991) J Biomol Struct Dyn 9:475–488). At the heart of these programs is a set of subroutines that, given the position of every atom in the model, calculate the total potential energy of the system and the force on each atom. These programs may utilize a starting set of atomic coordinates, such as the model coordinates provided in crystallographic-derived models, the parameters for the various terms of the potential energy function, and a description of the molecular topology (the covalent structure). Common features of such molecular modeling methods include: provisions for handling hydrogen bonds and other constraint forces; the use of periodic, boundary conditions; and provisions for occasionally adjusting positions, velocities, or other parameters in order to maintain or change temperature, pressure, volume, forces of constraint, or other externally controlled conditions.
Most conventional energy minimization methods use the input data described above and the fact that the potential energy function is an explicit, differentiable function of Cartesian coordinates, to calculate the potential energy and its gradient (which gives the force on each atom) for any set of atomic positions. This information can be used to generate a new set of coordinates in an effort to reduce the total potential energy and, by repeating this process over and over, to optimize the molecular structure under a given set of external conditions. These energy minimization methods are routinely applied to molecules similar to the subject hedgehog proteins as well as nucleic acids, polymers and zeolites.
In general, energy minimization methods can be carried out for a given temperature, Ti, which may be different than the docking simulation temperature, To. Upon energy minimization of the molecule at Ti, coordinates and velocities of all the atoms in the system are computed. Additionally, the normal modes of the system are calculated. It will be appreciated by those skilled in the art that each normal mode is a collective, periodic motion, with all parts of the system moving in phase with each other, and that the motion of the molecule is the superposition of all normal modes. For a given temperature, the mean square amplitude of motion in a particular mode is inversely proportional to the effective force constant for that mode, so that the motion of the molecule will often be dominated by the low frequency vibrations.
After the molecular model has been energy minimized at Ti, the system is “heated” or “cooled” to the simulation temperature, To, by carrying out an equilibration run where the velocities of the atoms are scaled in a step-wise manner until the desired temperature, TO, is reached. The system is further equilibrated for a specified period of time until certain properties of the system, such as average kinetic energy, remain constant. The coordinates and velocities of each atom are then obtained from the equilibrated system.
Further energy minimization routines can also be carried out. For example, a second class of methods involves calculating approximate solutions to the constrained EOM for the protein. These methods use an iterative approach to solve for the Lagrange multipliers and, typically, only need a few iterations if the corrections required are small. The most popular method of this type, SHAKE (Ryckaert et al. (1977) J Comput Phys 23:327; and Van Gunsteren et al. (1977) Mol Phys 34:1311) is easy to implement and scales as O(N) as the number of constraints increases. Therefore, the method is applicable to macromolecules such as the Hedgehog proteins of the present invention. An alternative method, RATTLE (Anderson (1983) J Comput Phys 52:24) is based on the velocity version of the Verlet algorithm. Like SHAKE, RATTLE is an iterative algorithm and can be used to energy minimize the model of the subject hedgehog protein.
The increasing availability of biomacromolecule strictures of potential pharmacophoric molecules that have been solved crystallographically has prompted the development of a variety of direct computational methods for molecular design, in which the steric and electronic properties of catalytic and substrate recognition sites are use to guide the design of potential inhibitors (Cohen et al. (1990) J. Med. Cam. 33: 883–894; Kuntz et al. (1982) J. Mol. Biol 161: 269–288; DesJarlais (1988) J. Med. Cam. 31: 722–729; Bartlett et al. (1989) (Spec. Publ., Roy. Soc. Chem.) 78: 182–196Goodford et al. (1985) J. Med. Cam. 28: 849–857; DesJarlais et al. J. Med. Cam. 29: 2149–2153). Directed methods generally fall into two categories: (1) design by analogy in which 3-D structures of known molecules (such as from a crystallographic database) are docked to the enzyme structure and scored for goodness-of-fit; and (2) de novo design, in which the ligand model is constructed-piece-wise in the enzyme. The latter approach, in particular, can facilitate the development of novel molecules, uniquely designed to bind to, and, e.g., inhibit the proteolytic activity of a hedgehog protein.
In an illustrative embodiment, the design of potential hedgehog inhibitors begins from the general perspective of shape complimentary for the active site and substrate specificity subsites of the enzyme, and a search algorithm is employed which is capable of scanning a database of small molecules of known three-dimensional structure for candidates which fit geometrically into the target protein site. It is not expected that the molecules found in the shape search will necessarily be leads themselves, since no evaluation of chemical interaction necessarily be made during the initial search. Rather, it is anticipated that such candidates might act as the framework for further design, providing molecular skeletons to which appropriate atomic replacements can be made. Of course, the chemical complimentary of these molecules can be evaluated, but it is expected that atom types will be changed to maximize the electrostatic, hydrogen bonding, and hydrophobic interactions with the enzyme. Most algorithms of this type provide a method for finding a wide assortment of chemical structures that are complementary to the shape of a binding site of the subject enzyme. Each of a set of small molecules from a particular data-base, such as the Cambridge Crystallographic Data Bank (CCDB) (Allen et al. (1973) J. Chem. Doc. 13: 119), is individually docked to the binding site of the hedgehog proteolytic domain in a number of geometrically permissible orientations with use of a docking algorithm. In a preferred embodiment, a set of computer algorithms called DOCK, can be used to characterize the shape of invaginations and grooves that form the active sites and recognition surfaces of the subject protein (Kuntz et al. (1982) J. Mol. Biol 161: 269–288). The program can also search a database of small molecules for templates whose shapes are complementary to particular binding sites of the enzyme (DesJarlais et al. (1988) J Med Chem 31: 722–729). These templates normally require modification to achieve good chemical and electrostatic interactions (DesJarlais et al. (1989) ACS Symp Ser 413: 60–69). However, the program has been shown to position accurately known cofactors for inhibitors based on shape constraints alone.
The orientations are evaluated for goodness-of-fit and the best are kept for further examination using molecular mechanics programs, such as AMBER or CHARMM. Such algorithms have previously proven successful in finding a variety of molecules that are complementary in shape to a given binding site of a receptor-enzyme, and have been shown to have several attractive features. First, such algorithms can retrieve a remarkable diversity of molecular architectures. Second, the best structures have, in previous applications to other proteins, demonstrated impressive shape complementarity over an extended surface area. Third, the overall approach appears to be quite robust with respect to small uncertainties in positioning of the candidate atoms.
Goodford (1985, J Med Chem 28:849–857) and Boobbyer et al. (1989, J Med Chem 32:1083–1094) have produced a computer program (GRID) which seeks to determine regions of high affinity for different chemical groups (termed probes) on the molecular surface of the binding site. GRID hence provides a tool for suggesting modifications to known ligands that might enhance binding. It may be anticipated that some of the sites discerned by GRID as regions of high affinity correspond to “pharmacophoric patterns” determined inferentially from a series of known ligands. As used herein, a pharmacophoric pattern is a geometric arrangement of features of the anticipated ligand that is believed to be important for binding. Attempts have been made to use pharmacophoric patterns as a search screen for novel ligands (Jakes et al. (1987) J Mol Graph 5:41–48; Brint et al. (1987) J Mol Graph 5:49–56; Jakes et al. (1986) J Mol Graph 4:12–20); however, the constraint of steric and “chemical” fit in the putative (and possibly unknown) receptor binding site is ignored. Goodsell and Olson (1990, Proteins: Struct Funct Genet 8:195–202) have used the Metropolis (simulated annealing) algorithm to dock a single known ligand into a target protein. They allow torsional flexibility in the ligand and use GRID interaction energy maps as rapid lookup tables for computing approximate interaction energies. Given the large number of degrees of freedom available to the ligand, the Metropolis algorithm is time-consuming and is unsuited to searching a candidate database of a few thousand small molecules.
Yet a further embodiment of the present invention utilizes a computer algorithm such as CLIX which searches such databases as CCDB for small molecules which can be oriented in the receptor binding site in a way that is both sterically acceptable and has a high likelihood of achieving favorable chemical interactions between the candidate molecule and the surrounding amino acid residues. The method is based on characterizing the receptor site in terms of an ensemble of favorable binding positions for different chemical groups and then searching for orientations of the candidate molecules that cause maximum spatial coincidence of individual candidate chemical groups with members of the ensemble. The current availability of computer power dictates that a computer-based search for novel ligands follows a breadth-first strategy. A breadth-first strategy aims to reduce progressively the size of the potential candidate search space by the application of increasingly stringent criteria, as opposed to a depth-first strategy wherein a maximally detailed analysis of one candidate is performed before proceeding to the next. CLIX conforms to this strategy in that its analysis of binding is rudimentary—it seeks to satisfy the necessary conditions of steric fit and of having individual groups in “correct” places for bonding, without imposing the sufficient condition that favorable bonding interactions actually occur. A ranked “shortlist” of molecules, in their favored orientations, is produced which can then be examined on a molecule-by-molecule basis, using computer graphics and more sophisticated molecular modeling techniques. CLIX is also capable of suggesting changes to the substituent chemical groups of the candidate molecules that might enhance binding.
The algorithmic details of CLIX is described in Lawerence et al. (1992) Proteins 12:31–41, and the CLIX algorithm can be summarized as follows. The GRID program is used to determine discrete favorable interaction positions (termed target sites) in the binding site of the protein for a wide variety of representative chemical groups. For each candidate ligand in the CCDB an exhaustive attempt is made to make coincident, in a spatial sense in the binding site of the protein, a pair of the candidate's substituent chemical groups with a pair of corresponding favorable interaction sites proposed by GRID. All possible combinations of pairs of ligand groups with pairs of GRID sites are considered during this procedure. Upon locating such coincidence, the program rotates the candidate ligand about the two pairs of groups and checks for steric hindrance and coincidence of other candidate atomic groups with appropriate target sites. Particular candidate/orientation combinations that are good geometric fits in the binding site and show sufficient coincidence of atomic groups with GRID sites are retained.
Consistent with the breadth-first strategy, this approach involves simplifying assumptions. Rigid protein and small molecule geometry is maintained throughout. As a first approximation rigid geometry is acceptable as the energy minimized coordinates of the hedgehog deduced structure, describe an energy minimum for the molecule, albeit a local one. If the surface residues of the site of interest are not involved in crystal contacts then the crystal configuration of those residues. We believe that the deduced crystal structure described in herein should reasonably mimic the mean solution configuration. Moreover, the equivalent models of of hedgehog isoforms (Ihh, Dhh, etc) can be derived by the same method.
A further assumption implicit in CLIX is that the potential ligand, when introduced into the active site of hedgehog protein, does not induce change in the protein's stereochemistry or partial charge distribution and so alter the basis on which the GRID interaction energy maps were computed. It must also be stressed that the interaction sites predicted by GRID are used in a positional and type sense only, i.e., when a candidate atomic group is placed at a site predicted as favorable by GRID, no check is made to ensure that the bond geometry, the state of protonation, or the partial charge distribution favors a strong interaction between the protein and that group. Such detailed analysis should form part of more advanced modeling of candidates identified in the CLIX shortlist.
Yet another embodiment of a computer-assisted molecular design method for identifying inhibitors of the subject hedgehog protein comprises the de novo synthesis of potential inhibitors by algorithmic connection of small molecular fragments that will exhibit the desired structural and electrostatic complementarity with the active site of the enzyme. The methodology employs a large template set of small molecules with are iteratively pieced together in a model of the hedgehog active site. Each stage of ligand growth is evaluated according to a molecular mechanics-based energy function, which considers van der Waals and coulombic interactions, internal strain energy of the lengthening ligand, and desolvation of both ligand and enzyme. The search space can be managed by use of a data tree which is kept under control by pruning according to the binding criteria.
In an illustrative embodiment, the search space is limited to consider only amino acids and amino acid analogs as the molecular building blocks. Such a methodology generally employs a large template set of amino acid conformations, though need not be restricted to just the 20 natural amino acids, as it can easily be extended to include other related fragments of interest to the medicinal chemist, e.g. amino acid analogs. The putative ligands that result from this construction method are peptides and peptide-like compounds rather than the small organic molecules that are typically the goal of drug design research. The appeal of the peptide building approach is not that peptides are preferable to organics as potential pharmaceutical agents, but rather that: (1) they can be generated relatively rapidly de novo; (2) their energetics can be studied by well-parameterized force field methods; (3) they are much easier to synthesize than are most organics; and (4) they can be used in a variety of ways, for peptidomimetic inhibitor design, protein-protein binding studies, and even as shape templates in the more commonly used 3D organic database search approach described above.
Such a de novo peptide design method has been incorporated in a software package called GROW (Moon et al. (1991) Proteins 11:314–328). In a typical design session, standard interactive graphical modeling methods are employed to define the structural environment in which GROW is to operate. For instance, environment could be the active site cleft of hedgehog, or it could be a set of features on the protein's surface to which the user wishes to bind a peptide-like molecule, a peptide sequence based on the cleavage site of hedgehog itself (e.g., to represent the autoproteolytic event). The GROW program then operates to generate a set of potential ligand molecules. Interactive modeling methods then come into play again, for examination of the resulting molecules, and for selection of one or more of them for further refinement.
To illustrate, GROW operates on an atomic coordinate file generated by the user in the interactive modeling session, such as the coordinates provided in the crystallographic-derived models, plus a small fragment (e.g., an acetyl group) positioned in the active site to provide a starting point for peptide growth. These are referred to as “site” atoms and “seed” atoms, respectively. A second file provided by the user contains a number of control parameters to guide the peptide growth (Moon et al. (1991) Proteins 11:314–328).
The operation of the GROW algorithm is conceptually fairly simple. GROW proceeds in an iterative fashion, to systematically attach to the seed fragment each amino acid template in a large preconstructed library of amino acid conformations. When a template has been attached, it is scored for goodness-of-fit to the receptor site, and then the next template in the library is attached to the seed. After all the templates have been tested, only the highest scoring ones are retained for the next level of growth. This procedure is repeated for the second growth level; each library template is attached in turn to each of the bonded seed/amino acid molecules that were retained from the first step, and is then scored. Again, only the best of the bonded seed/dipeptide molecules that result are retained for the third level of growth. The growth of peptides can proceed in the N-to-C direction only, the reverse direction only, or in alternating directions, depending on the initial control specifications supplied by the user. Successive growth levels therefore generate peptides that are lengthened by one residue. The procedure terminates when the user-defined peptide length has been reached, at which point the user can select from the constructed peptides those to be studied further. The resulting data provided by the GROW procedure include not only residue sequences and scores, but also atomic coordinates of the peptides, related directly to the coordinate system of the receptor site atoms.
In yet another embodiment, potential pharmacophoric compounds can be determined using a method based on an energy minimization-quenched molecular dynamics algorithm for determining energetically favorable positions of functional groups in the binding sites of the subject hedgehog protein. The method can aid in the design of molecules that incorporate such functional groups by modification of known ligands or de novo construction.
For example, the multiple copy simultaneous search method (MCSS) described by Miranker et al. (1991) Proteins 11: 29–34. To determine and characterize a local minima of a functional group in the forcefield of the protein, multiple copies of selected functional groups are first distributed in a binding site of interest on the hedgehog protein. Energy minimization of these copies by molecular mechanics or quenched dynamics yields the distinct local minima. The neighborhood of these minima can then be explored by a grid search or by constrained minimization. In one embodiment, the MCSS method uses the classical time dependent Hartee (TDH) approximation to simultaneously minimize or quench many identical groups in the forcefield of the protein.
Implementation of the MCSS algorithm requires a choice of functional groups and a molecular mechanics model for each of them. Groups must be simple enough to be easily characterized and manipulated (3–6 atoms, few or no dihedral degrees of freedom), yet complex enough to approximate the steric and electrostatic interactions that the functional group would have in binding to the site of interest in the hedgehog protein. A preferred set is, for example, one in which most organic molecules can be described as a collection of such groups (Patai's Guide to the Chemistry of Functional Groups, ed. S. Patai (New York: John Wiley, and Sons, (1989)). This includes fragments such as acetonitrile, methanol, acetate, methyl ammonium, dimethyl ether, methane, and acetaldehyde.
Determination of the local energy minima in the binding site requires that many starting positions be sampled. This can be achieved by distributing, for example, 1,000–5,000 groups at random inside a sphere centered on the binding site; only the space not occupied by the protein needs to be considered. If the interaction energy of a particular group at a certain location with the protein is Core positive than a given cut-off (e.g. 5.0 kcal/mole) the group is discarded from that site. Given the set of starting positions, all the fragments are minimized simultaneously by use of the TDH approximation (Elber et al. (1990) J Am Chem Soc 112: 9161–9175). In this method, the forces on each fragment consist of its internal forces and those due to the protein. The essential element of this method is that the interactions between the fragments are omitted and the forces on the protein are normalized to those due to a single fragment. In this way simultaneous minimization or dynamics of any number of functional groups in the field of a single protein can be performed. Minimization is performed successively on subsets of, e.g. 100, of the randomly placed groups. After a certain number of step intervals, such as 1,000 intervals, the results can be examined to eliminate groups converging to the same minimum. This process is repeated until minimization is complete (e.g. RMS gradient of 0.01 kcal/mole/Å). Thus the resulting energy minimized set of molecules comprises what amounts to a set of disconnected fragments in three dimensions representing potential pharmacophores.
The next step then is to connect the pharmacophoric pieces with spacers assembled from small chemical entities (atoms, chains, or ring moieties). In a preferred embodiment, each of the disconnected can be linked in space to generate a single molecule using such computer programs as, for example, NEWLEAD (Tschinke et al. (1993) J Med Chem 36: 3863,3870). The procedure adopted by NEWLEAD executes the following sequence of commands (1) connect two isolated moieties, (2) retain the intermediate solutions for further processing, (3) repeat the above steps for each of the intermediate solutions until no disconnected units are found, and (4) output the final solutions, each of which is single molecule. Such a program can use for example, three types of spacers: library spacers, single-atom spacers, and fuse-ring spacers. The library spacers are optimized structures of small molecules such as ethylene, benzene and methylamide. The output produced by programs such as NEWLEAD consist of a set of molecules containing the original fragments now connected by spacers. The atoms belonging to the input fragments maintain their original orientations in space. The molecules are chemically plausible because of the simple makeup of the spacers and functional groups, and energetically acceptable because of the rejection of solutions with van-der Waals radii violations.
The three-dimensional structure of hedgehog is useful to aid in screening and development of diagnostic and therapeutic protein fragments as in rational drug design, to search for structural analogs of known protein structures, or to aid in an analysis of biological function and activity. Also, the method may be used to predict protein secondary structures and protein subsecondary structures from amino acid sequences alone, and to predict those regions of a protein molecule that are on the outside and those that are on the inside.
Compounds can also be prepared using the three-dimensional structure provided herein and tested using assays known to those of skill in the art. For example, compounds can be synthesized and screened for hedgehog autoproteolytic activity by cleavage assays (see for example, Porter et al., Cell 86:21, 1996; WO96/17924, herein incorporated by reference).
Compounds of the invention include drugs, small molecules, peptides, peptidomimetics, polypeptides, chemical compounds and biologic agents. For example, peptidomimetics are synthetic compounds having a three-dimensional structure (i.e., a “peptide motif”) based upon the three-dimensional structure of a selected peptide. The peptide motif provides the peptidomimetic compound with Hedgehog agonist or antagonist activity that is substantially the same as, or greater than, the Hedgehog agonist or antagonist activity of the peptide from which the peptido-mimetic was derived. Peptidomimetic compounds can have additional characteristics that enhance their therapeutic application, e.g., enhanced cell permeability, increased receptor or polypeptide binding affinity and/or avidity, and prolonged biological half-life. The design of peptidomimetic compounds having agonist or antagonist activity can be aided through computer modeling techniques well known in the art. Other methods for the design, as well as the preparation of, peptidomimetic compounds are well known in the art.
Atomic coordinates and structure factors have been deposited in the Brookhaven Protein Data Bank. Applicant assures complete access and disclosure of these coordinates and factors upon issuance of a patent.
The following examples are intended to illustrate but not limit the invention. While they are typical of those that might be used, other procedures known to those skilled in the art may alternatively be used.
The full length form of the hh protein (F) migrates with a mobility corresponding to a relative molecular mass of 46 kD.
Immunoblots were performed with affinity purified Ab1 or Ab2 by either of two chemiluminescence based protocols. In the first protocol (used in
cDNAs encoding various hh protein species were cloned into the pMK33 vector, which allows for inducible expression under metallothionein promoter control (M. R. Koelle et al., Cell 67:59,1991). Stable S2 cell lines were made by transfection of the hh/pMK33 plasmids with constant selection for hygromycin resistance. Proteins were expressed by plating a log phase culture of cells diluted to 0.1 A595 units, waiting 48 hours, inducing with CUSO4 at 0.2 mM final concentration, and harvesting the cells and/or supernatant 24 hours later. Cell samples for immunoblotting were made by adding 10 volumes of 1× SDS PAGE loading buffer to pelleted cells.
In vitro translations were performed with the use of the TNT coupled transcription-translation system (Promega). 35S methionine (DuPont NEN) was used for detection by autoradiography. In the heparin binding experiment in vitro translation lysate with microsomes that produce wild-type hh protein was added to heparin agarose (Sigma) or Sepharose CL-4B (Pharmacia) beads pre-equilibrated with heparin binding buffer (HBB; 20 mM Tris (7.4), 150 mM NaCl, 0.1 percent Triton X-100). Samples were incubated at 4° C. for four hours with gentle rocking. After pelleting the beads, supernatants in some samples were analyzed (lanes 2 and 4). The beads were then washed 5 times with chilled HBB and samples (lanes 3 and 5) were subsequently eluted at 80° C. for 10 minutes in SDS PAGE loading buffer (F. M. Ausubel et al., supra).
Embryos from the wild-type Canton-S line and from the matings, hshh/hshh or hshh H329A-/hshh H329A X y; Sco/CyO, enlacZ11::wg (Kassis, et al., Proc. Natl. Acad. Sci. U.S.A. 89: 1919, 1992), were collected 0 to 16 hours after egg laying (AEL) at 25° C. They were heat shocked for 30 minutes at 37° C. and allowed to recover for 1 hour at 25° C. Embryos in
Multiple species were detected and minor cross reactive bands are seen in most samples including extracts of induced untransfected S2 cells (lanes 1 and 13). One of these bands (occurring in both panels) co-migrates with U (at 39 kD) and is particularly abundant in lane 6 of
Smaller species of Hh proteins from in vivo sources have been reported previously (T. Tabata and T. B. Kornberg, Cell 76: 89, 1994). The latter study examined not endogenous proteins, but proteins induced to express at high levels from exogenously introduced constructs. The antibody used did not distinguish epitopes from distinct portions of the molecule.
In addition to signal cleavage, a further cleavage of the U precursor is responsible for generating other forms of hh protein observed in vivo. This was deduced from the observation that Ab1 and Ab2 both detected the U (uncleaved) species, but also interacted individually with smaller protein species expressed endogenously in embryos and imaginal discs or with species expressed upon introduction of the hh gene into S2 cells. Ab1 thus interacts with a 19 kD species from all of these tissues (
The proposed cleavages by which these species arise are shown schematically in the bottom portion of
A simple mechanism that could account for the derivation of the two smaller species therefore would be a single internal cleavage of the U precursor. Processing of the hh protein when translated in vitro also yields a 25 kD species (C; lanes 16 and 17) and either a 29 kD or 19 kD (N) species (lanes 4 and 5). The 19 kD species comigrates with N, and its formation depends upon the presence of microsomes, consistent with the proposal that N derives from F by signal cleavage and a further internal cleavage. The overall pathway for formation of the predominant forms of hh protein observed in vivo thus appears to involve signal cleavage of F to generate U. U is then cleaved internally to form N and C, which are the predominant forms found in vivo. Further processing of the 25 kD C species might then generate the 16 kD C* species, but whether this processing is a single cleavage event or not is not clear since Ab2 does not recognize the smaller 9 kD fragment that would result. The processing of C to generate C* appears to occur with greater efficiency in imaginal discs as compared to embryos (compare lanes 15 and 18); this may be caused by the more extended mass isolation procedure of imaginal discs (O. M. Eugene, et al., Tissue Culture Assn. Man., 5: 1055, 1979).
The comigration of endogenous and in vitro-generated hh protein species suggested that in vitro processing is similar to that observed in vivo.
To determine whether this invariant His residue in the hh protein indeed plays a role in auto-proteolysis, two proteins from E. coli were purified: one carried the wild type sequence and the other a substitution of an Ala codon for the His codon at position 329 (H329A). Both of these proteins were engineered to contain a hexa-histidine tag at the amino terminus fused to Drosophila sequences extending from a residue just before the signal cleavage site to the carboxy-terminus (residues 83 to 471; the wild type form of this protein is referred to as His6-U). Both proteins were extensively purified under denaturing conditions using a Ni++-chelating matrix.
Preliminary proteinase inhibitor studies have been performed on in vitro translated Hh protein by adding various inhibitors at the start of the translation reaction. These studies have been complicated by the fact that numerous protease inhibitors lower or block translation efficiency. In some cases the effectiveness of an inhibitor was assayed by determining if addition of an inhibitor to a completed translation reaction will inhibit the self-processing that normally continues to occur. At this time we can only state the following with certainty: (i) the serene protease inhibitor TAME (p-toluenesulfonyl-L-arginine methyl ester) inhibits auto-proteolysis of in-vitro translated Hh protein; (ii) soybean trypsin inhibitor, a, anti-trypsin, aprotitin, leupeptin, and E-64 do not block auto-proteolysis of translated Hh protein; and (iii) TAME partially inhibits auto-proteolysis of purified His6-U protein (
As seen in
To more precisely define the domain of the hh protein responsible for this auto-proteolytic event, the effects of several distinct types of mutations upon in vitro processing were examined. The most informative mutation was a deletion that removes residues 89 to 254 (Δ89–254), which together constitute most of the amino acids within the portion of the molecule presumed to form the N fragment. In vitro translations of wild-type and mutant Hh proteins from Drosophila (
In vitro translations were performed with the use of the TNT coupled transcription-translation system (Promega). 35S methionine (DuPont NEN) was used for detection by autoradiography. In the heparin binding experiment (
All mutations in the hh gene were generated in the plasmid pF1 (J. J. Lee, et al., supra). Mutations in the zebrafish twhh and shh genes were generated with the original cDNA clones as described (Ekker, et al., Current Biology, 5(U8: 944,1995). All point mutations were generated with the use of recombinant circle PCR (D. H. Jones and S. C. Winistorfer, Biotechniques 12: 528, 1992). The flu408 and flu227 mutations were generated by inserting a trimer of the influenza hemagglutinin antigen (42 residues for flu408 and 43 residues for flu227) into the AlwN I and Bgl I sites present in the hh ORF (nucleotide positions 1604 and 1058 respectively) (J. J. Lee. et al., supra). The Δ89–254 mutation was generated by removing sequences between the EcoN I site (644) and the Pml I site (1145). The 294 trunc mutation was generated by removing sequences between the Acc I site (1265) and the Xcm I site (1792). The 410 trunc mutation was previously generated and identified as Hh410 (J. J. Lee, et al., supra). To map the mutation in the hh13E allele (base change C1756 to A; coding change Tyr457 to STOP), DNA isolated from hh13E/TM3 was used to seed PCR reactions generating regions of the hh ORF and flanking sequences, which were subcloned into Bluescript KSM (Stratagene). Six clones each, derived from two different PCR amplifications were sequenced.
As seen in lanes 1 and 2 of
In contrast, lesions affecting residues presumed to lie within C block auto-proteolysis in vitro. All mutations tested by in vitro translation were also examined in S2 cells by immunoblotting. In all cases the patterns of cleavage in S2 cells were identical to those observed in translations except that C* was always present whenever C was formed. The former fragment was not observed in translations. These include the H329A mutation described above, a mutation that inserts an influenza virus epitope between residues 408 and 409 (flu408), and three mutations that cause premature termination of the protein at the carboxy terminus. The two most severe truncations, 294 trunc and 410 trunc, are mutations generated in vitro. They cause a loss of 177 and 61 residues, respectively, from the carboxyl-terminus of the protein, and neither undergoes proteolysis. The 456 trunc hh protein is like that encoded by the EMS-induced hh13E mutant allele, which results in the loss of 15 residues from the carboxy-terminus of the protein. This protein undergoes auto-proteolysis, as demonstrated by the appearance of a 24 kD band in place of C, but the efficiency of the reaction is much impaired in vitro (
The sequence homology and auto-proteolytic function of the full length hh protein suggested the possibility that F or the C fragment is a sequence-specific protease. As a first step in clarifying the mechanism of auto-proteolysis, an influenza virus epitope tag was introduced into the N-terminus of a hh open reading frame that also carried a H329A mutation.
The hh gene has been broadly conserved in evolution, with single homologues unidentified in a wide variety of invertebrate species and multiple distinct homologues in each of several vertebrate species (Y. Echelard et al., Cell 75: 1417, 1993; S. Krauss, et al., Cell 75: 1431, 1993; H. Roelink et al., Cell, supra). As seen in
Numerous functions for the hh gene have been described in Drosophila. At the morphological level these include a role in patterning of larval cuticular structures and adult structures such as the eye and appendages (C. Nüisslein-Volhard and E. Wieschaus, Nature 287: 795, 1980; and J. Mohler, Genetics 120: 1061, 1988).; the mechanistic basis for thee morphological effects involves signaling for maintenance or induction of gene expression in embryos and imaginal discs (J. J. Lee, supra; T. Tabata and T. B. Komberg, Cell 76: 89, 1994; and K. Basler and G. Struhl, Nature 368: 208, 1994). To ascertain the importance of auto-proteolysis for these functions, the H329A mutant gene under control of the hsp 70 promoter was introduced by P element-mediated transformation into the Drosophila germline. The hshh H329A construct was made identically to the hshh construct with the use of a hh ORF fragment containing the H329A mutation. Transgenic flies were generated from a y1 w1118 parental strain using standard methods of P element mediated transformation (A. C. Spradling and G. M. Rubin, Science 218: 341 1982). A line, HA3, carrying the hshh H329A P element on the second chromosome was maintained as a homozygous stock. To assay for expansion of wg stripes, embryos collected at 4 to 6 hours after egg laying (AEL) at 25° C. were subjected to the following heat shock protocols prior to fixation. Embryos receiving single shocks (10 or 30 minutes at 37° C.) were allowed to recover for 1 hour at 25° C. Embryos receiving double shocks (two 10 minute or two 30 minute shocks at 37° C.) were allowed to recover 90 minutes after the first shock and 40 minutes after the second (Both recoveries were at 25° C. The double 30 minute protocol was as previously described, (S. Krauss, supra). In situ hybridizations were performed as described (D. Tautz, Chromosoma 98: 81, 1989) using a wg specific probe (D. T. Chang et al., supra). Embryos assayed for cuticle phenotype were heat shocked 6 to 8 hours AEL for 30 minutes at 37° C., allowed to develop at 25° C. for 36 hours and then processed and mounted as described (M. Ashburner, Drosophila: A Laboratory Manual, Cold Spring Harbor Laboratory Press, New York, 1989). Immunolocalizations (single or double stains) were performed as described. With the use of affinity purified Ab1 or Ab2 for the primary antibody and alkaline phosphatase (AP) or horseradish peroxidase (BRP) conjugated anti rabbit or mouse IgG (Jackson Immuno-research) for the secondary. Embryos from a hh13E/TM3ftz-lacZ (the balancer chromosome was from the Bloomington Stock Center, strain 3218) stock homozygous for the hh13E allele were identified by the lack of staining with an anti b-galactosidase antibody (Promega), in a double stain with Ab2 (
Perhaps the earliest known requirement for Hh protein is in maintenance of an adjacent stripe of wingless (wg) gene expression in each embryonic segment (A. Martinez Arias, et al., Development 103: 157, 1988; and S. DiNardo, et al., Nature 332: 604, 1988). This requirement is deduced from the loss of wg expression when hh function is absent; in addition, the ubiquitous expression of wild-type Hh protein induces expansion of the domain of wg gene expression (P. W. Ingham, Nature 366: 560, 1993). The effects of the H329A mutation upon wg expansion were examined by heat shocking embryos carrying the H329A mutant construct in parallel with embryos containing the wild-type construct. Although the H329A mutant protein is able to induce some expansion of the wg domain, the efficiency of this activity is impaired relative to that of the wild-type protein (
The effects of Hh protein on the patterning of cuticular structures are most clearly visible on the dorsal surface of the larva, where four distinctive cell types can be identified in each para-segment. These cell types have been designated 1°, 2°, 3°, and 4°, from anterior to posterior, with hh transcription occurring in precursors of the 1° cells (J. Heemskerk and S. DiNardo, supra). Differentiation of the first three cell types was shown to be dependent upon hh gene function, and it has been proposed that the fates of these cells are determined by the concentration of Hh protein, with highest concentrations producing the 1° fate, intermediate concentrations producing the 2° fate, and the lowest concentrations producing the 3° fate (J. Heemskerk and S. DiNardo, supra). This proposal was supported by observations that the most anterior cell types display the greatest sensitivity to a reduction of hh expression, and that all of the 3° and some of the 4° bristles are replaced by naked cuticle characteristic of the more anterior 2° cell type when hh is expressed ubiquitously at high levels. We have reproduced suppression 3° and some 4° fates by heat shock induction of embryos that carry our wild-type construct (
Studies of H329A mutant protein were extended to the function to the patterning of adult structures and signaling within imaginal discs. In the eye imaginal disc hh function is required for appropriate development of pattern (J. Mohler, Genetics 120: 1061, 1988; J. J. Lee, supra; and J. Mohler. and K. Vani, supra) and more recently has been shown to control progression of a wave of differentiation via induction of decapentaplegic (dpp) gene expression in the morphogenetic furrow of the eye (U. Heberlein, et al., Cell 75: 913, 1993; and C. Ma, et al., Cell 75: 927, 1993). In leg and wing discs, ectopic expression of hh has also been shown to yield pattern duplications and defects and is associated with induction of ectopic expression of other signaling molecules normally expressed in a zone along the anterior/posterior compartment boundary (T. Tabata and T. B. Kornberg, Cell 76: 89, 1994; and K. Basler and G. Struhl, Nature 368: 208, 1994).
For studies of signaling in imaginal discs, a thermal cycler was utilized to subject larvae carrying heat shock-inducible hh constructs to successive rounds of heat shock and recovery. The effects of temperature cycling upon expression of dpp and wg in imaginal discs was examined by monitoring β-galactosidase expression from a reporter gene carrying dpp promoter sequences or from an enhancer detector P element inserted in the wg gene. In
Virgin female flies from the homozygous lines hshh (D. T. Chang et al., Development, 1994, in press), hshh H329A, and y1 w1118 were crossed to males from the homozygous BS3.0 line (bearing a P element dpp reporter construct on the 2nd chromosome, referred to as dpp-lacZ) (. K. Blackman, et al., Development 111: 657, 1991) or the line y; Sco/ CyO, enlacZ11::wg (bearing a wg reporter P element enhancer trap on a second chromosome balancer; called wg-lacZ) (J. A. Kassis, et al., Proc. Natl. Acad. Sci. U.S.A. 89: 1919, 1992). Progeny were grown at 25° C. in aerated 0.5-ml microcentrifuge tubes containing yeast paste until the late second instar or early third instar stage of larval development. The larvae were then cycled continuously at 37° C for 30 minutes followed by 25° C. for 90 minutes in a Perkin-Elmer thermal cycler until they reached the late third instar stage. They were subsequently dissected and stained with X-gal as described (M. Ashburner, supra) or allowed to grow to adulthood for phenotypic analysis.
As shown in
Ubiquitous expression of wild-type hh also leads to ectopic expression of dpp in the eye-antennal disc (compare
The experiments described thus far comprise multiple series of larvae subjected to two days of thermal cycling followed by immediate dissection for analysis of imaginal structures or further incubation at constant temperature for analysis of adult structures. Although the H329A protein appeared to have little activity in these experiments, the small patch of ectopic dpp expression induced in the wing disc (
A puzzling feature of hh function is its apparent short-range action in settings such as embryonic and imaginal disc signaling to wg and dpp, and longer-range action in other settings, such as patterning of the dorsal larval cuticle. These observations and the existence of two major protein products in vivo prompted us to look for differences in the solubility or diffusibility of N and C expressed in S2 cultured cells.
One possible explanation for this differential behavior might be association of the N fragment with extracellular matrix proteins on the surfaces of the S2 cells. Accordingly, the relative affinity of these two proteins for heparin agarose was examined, since heparin binding is a common property of proteins that associate with the extracellular matrix. Given the obvious difficulty in obtaining soluble N from cultured cells, in vitro translation in the presence of microsomes was used to generate soluble, labelled N and C. As shown in
The differential release of N and C into cultured cell supernatants suggested the possibility that these fragments might also be differentially localized in embryos. Previously reported hh protein localizations utilized either antibodies specific for N epitopes or antibodies unable to distinguish between N and C.
The hh13E mutation encodes a prematurely truncated protein that is missing 15 residues normally present at the COOH-terminus. Because this protein displays a much reduced efficiency in auto-proteolysis the distribution of C in this mutant background was examined.
In addition to signal cleavage, the hh protein undergoes auto-proteolysis at an internal site to generate the predominant protein species observed in vivo. All or most of the amino acid residues required for this auto-proteolysis function map to C, the carboxy-terminal product of this internal cleavage. In an effort to determine the importance of auto-proteolysis for function, we introduced a single residue mutation (H329A) that blocks auto-proteolysis of the hh protein in vitro and demonstrated that both processing and function of this protein is impaired in vivo. Since similar levels of induced protein were detected from a strain carrying the wild-type construct or from several strains carrying independent insertions of the mutant construct (
Curiously, the H329A Hh protein appears to retain weak activity in embryonic signaling to induce ectopic wg expression and, to a lesser degree, can function in imaginal disc signaling for induction of ectopic wg and dpp expression. In contrast to its retention of at least some signaling functions in embryonic and imaginal tissues, the H329 protein is completely inert when assayed for the ability to reprogram cell fates in the dorsal cuticle of the larva.
The assays in which the H329A protein is active or partially active involve short-range signaling that normally occurs across one or at most several cell diameters; in contrast, the H329A protein fails to exert any effect upon patterning of the dorsal cuticle, a long-range activity that normally operates across most of the segment. Previous proposals to account for long-range patterning activities have suggested that hh expression induces other signaling molecules which are then responsible for executing the patterning functions (the signal relay model; see
These proposals seek to maintain a consistent mode of hedgehog action by rationalizing the apparent long-range activities of hh products as indirect consequences of short-range signaling. Based on the distribution observed, the active molecule in this model might be N and the role of C would then be limited to supplying the catalytic machinery required for biogenesis of N.
Our evidence suggests an alternative model, the dual function model (
Execution of short-range functions by N would be consistent with the observation that the H329A mutant protein has at least partial function in signaling for the induction of wg and dpp, since this mutation does not alter residues located in the amino-terminal portion of the protein that normally would give rise to N. The uncleaved H329A protein thus would carry all the residues that normally interact with a presumed receptor for N, although there might be some effect on the affinity of the interaction due to the presence of carboxy-terminal sequences, thus accounting for the decreased potency of the H329A protein. Alternatively, the partial function of H329A protein may derive from an extremely small fraction of protein that appears to be cleaved, a very faint band with identical mobility to C appears in in vitro translations with the H329A protein (
When N is synthesized from a native construct (wild type hh), it remains primarily cell-associated (
The mouse and human hh-like sequences were isolated by polymerase chain reaction (PCR) using primers degenerate for all possible coding combinations of the sequences underlined in
Mouse clones obtained in this manner contained 144 bases of sequence between the primer ends and were labelled with [α−32P]dATP and used for high stringency screens of mouse cDNA libraries made from whole 8.5 dpc embryonic RNA and from 14.5 dpc embryonic brain in the λZAP vector (a gift from A. Lanahan). Several clones corresponding to Hhg-1 were isolated and the largest, 2629 bp in length (pDTC8.0), was chosen for sequence analysis using dideoxy chain termination (Sanger, et al., 1977) and Sequenase v2.0 (US Biochemicals). Compressions were resolved by using 7-deaza- guanosine (US Biochemicals). Sequence analysis made use of the Geneworks 2.0 (IntelliGenetics) and MacVector 3.5 (IBI) software packages.
One of the three mouse clones, Hhg-1, when used as a probe, yielded a 2.0 kb clone from a 8.5 dpc mouse embryonic cDNA library and a 2.7 kb clone from a 14.5 dpc embryonic cDNA library. The 2.7 kb cDNA appears to represent a nearly full length mRNA because it corresponds to a 2.7 kb band detected by hybridization on a Northern blot. The largest methionine-initiated open reading frame within this cDNA encompasses 437 codons, and is preceded by one in frame upstream stop codon. Sequence comparisons indicate that the protein encoded by Hhg-1 is identical to the independently characterized mouse Shh (Echelard, et al., Cell, 75:1417–1430, 1993) except for an arginine to lysine difference at residue 122. Hhg-1 also corresponds closely to the rat vhh-1 gene (97% amino acid identity; Roelink, et al., Cell, 76:761–775, 1994), the chicken Sonic hedgehog (81% identity; Riddle, et al., Cell, 75:1401–1416, 1993) and Shh from the zebrafish (68% identity; Krauss, et al., Cell, 75:1431–1444, 1993; Roelink, et al., Cell, 76:761–775, 1994). The PCR-generated fragments Hhg-2 and Hhg-3 appear to correspond to the Indian and Desert classes of mouse hedgehog genes, respectively (Echelard, et al., Cell, 75:1417–1430, 1993).
Alignment of the Hhg-1 open reading frame with the two Drosophila hh sequences showed that all three proteins contain hydrophobic amino acid sequences near their amino-termini; the hydrophobic stretches within the D. melanogaster protein (residues 64 to 83) and within the mouse protein are known to act efficiently as signal sequences for cleavage (Lee, et al, Cell, 71:33–50, 1992). Both Drosophila signal sequences are unusual in their internal locations, while the hydrophobic stretch of the mouse gene occurs at the extreme amino-terminus, a more conventional location for cleaved signal sequences. Although portions of sequence N-terminal to the Drosophila signal sequences are conserved, suggesting a functional role, the mouse gene lacks this region.
The overall level of amino acid identity between Hhg-1 and hh carboxy-terminal to the signal sequences is 46%. A closer examination shows that the amino terminal portion, from residues 25 to 187, displays 69% identity, while remaining residues in the carboxy-terminal portion display a much lower 31% identity. Like hh, the Hhg-1 coding sequence is divided into three exons, and the boundaries of these exons are at the same positions within coding sequence as those of the three Drosophila hh exons. Curiously, the boundary between coding sequences of the second and third exons occurs near the transition from high to low levels of overall sequence conservation. The coincidence of these two boundaries suggests a possible demarcation of functional domains within these proteins. This location within Hhg-1 coding sequence also coincides approximately with the site of a presumed proteolytic cleavage.
Partial sequence for two human hh genes has been obtained by DNA sequencing of clones derived by PCR amplification from genomic DNA with hh-specific degenerate primers as outlined in Chang, et al., (Development, 120:3339, 1994) and EXAMPLE 9 (
For example, ready-made cDNA libraries or RNAs from a variety of human sources, including various fetal stages and organs (from abortuses) and specific infant or adult organs (from pathological or autopsy specimens), are being tested for the presence of hh sequences by PCR or RT-PCR using the primers described in Chang, et al., supra, and other primers derived directly from the sequence of the human fragments. Ready-made libraries containing hh sequences are being screened directly and, where necessary, new libraries are being constructed by standard methods from RNA sources containing hh sequences. The probe for these screens is a mixture of all the distinct human hh fragments. Sequences of cDNA clones can then be determined. Most clones containing the probe sequences, which are located in the N region, will also include a full C coding region since standard methods of library construction result in cDNA clones that are most complete at their 3′ ends. All full length hh-coding sequences obtained previously in vertebrates and invertebrates contain N and C sequences encoded in a single RNA. Screening is continued until complete open reading frames that correspond to all of the fragments of human hh genes are obtained. Specifically, 1.2×106 clones from a human fetal brain library (—Stratagene, La Jolla, Calif.) was screened using a mixture of the two human hh fragments (
Second, the RNA sources identified as containing hh sequences can be used as templates from anchored PCR (also referred to in the literature as RACE, for rapid amplification of cDNA ends). Briefly, this method provides a means to isolate further mRNA sequence in either the 5′ or 3′ direction provided that sequence is known from an internal starting point. Anchored PCR can also be used to isolate sequences from cDNA library.
Third, genomic libraries can be screened with the probes described in the first technique. Where necessary, human hh exons and coding sequences are being identified by hybridization to previously isolated human and mouse coding sequences by sequence determination, and by exon-trapping methods to identify all hh coding sequences within genomic clones; these coding sequences can be “stitched” together by standard recombinant DNA methods to generate complete hh open reading frames.
Partial sequences corresponding to five distinct zebrafish hh-like genes were isolated and the complete coding sequences for two of these genes were obtained from an embryonic cDNA library. One of these two sequences is identical to that of the zebrafish nhh-I gene (Roelink, et al., Cell, 76:761, 1994), and appears to correspond to the shh gene reported by Krauss, et al., (Cell, 75:1431, 1993) (See
The first four sequences were isolated from zebrafish genomic DNA (a gift from J. Pellegrino) using degenerate primers in polymerase chain reactions as described (Chang, et al., supra). twhh and shh clones were isolated from a 20–28 hour cDNA library (a gift from R. Riggleman, K. Helde, D. Grunwald and J. Pellegrino) using the first three sequences as probes. The translational reading frames for twhh and shh were closed 12 and 16 codons, respectively, upstream of the putative initiating methionine.
Comparison of twhh and shh expression patterns (Krauss, et al, supra), reveals that both gene are predominantly expressed in midline structures, albeit with notable differences in regard to timing, rostra-caudal extent, and tissue restriction. Expression of twhh is first detected during gastrulation in the dorsal mesoderm (
As somitogenesis proceeds, ventral midline expression of shh and twhh is reduced in most of the prospective forebrain, but remains strong in an anterior patch of midline cells within the floor of the prospective diencephalon (
Differences between twhh and shh expression are apparent from the beginning of gastrulation, since twhh RNA can be detected as early as the shield stage while shh is first detected later, at about 60% epiboly (not shown; (Krauss, et al., supra). In addition, twhh transcripts are restricted to neural tissues early in development, and are never detected in the notochord (compare
To gain insight into the potential roles of hh products in development, synthetic twhh and shh mRNA was injected into 1–8 cell embryos. This technique yields a mosaic but fairly uniform pattern of expression, as determined for the control mRNA encoding β-galactosidase (not shown). Uniformity of expression is in good agreement with fate mapping studies of the early zebrafish embryo (Kimmel & Warga, supra; Warga & Kimmel, supra; Heide, et al., Science, 265:517–520, 1994), which indicate that blastomeres undergo extensive cell mixing during the cleavages prior to gastrulation. We note that mosaicism of expression caused surprisingly little variation in the phenotypes of the hh injected embryos, possibly due to secretion of hh gene products.
Embryos injected with synthetic twhh or shh mRNA (hh RNA) exhibited numerous yet highly reproducible abnormalities in comparison to control embryos injected with lacZ mRNA. These abnormalities, discussed below, are primarily defects in the brain and eyes. Although the effects of ectopic twhh and shh expression were qualitatively similar, the incidence and severity were greater with twhh RNA (see text below,
Morphological defects in the brain and other rostral neural derivatives occur at high frequency in hh-injected embryos. The three ventricles of the fish brain normally apparent at 28 hours of development—the rhombencephalic, mesencephalic (
Defects in eye development also occur at high frequency in embryos injected with hh RNA. Thus, while at 28 hours the normal zebrafish eye has a lens and a retina with pigmented epithelium (
The eye phenotypes caused by hh RNA injection resemble those produced by treatment of zebrafish and Xenopus laevis embryos with retinoic acid. In Xenopus, phenotypes range from reduction of the eye and absence of the lends to eyes with retinal folds (resembling duplicated dyes) and multiple small lenses (Manns, M. & Fritzsch, B., Neurosci. Lett., 127:150–154, 1991). In zebrafish, exposure to retinoic acid during gastrulation interferes with the formation of the eye (Holder, N. & Hill, J., Development, 113:1159–1170, 1991), while exposure during formation of the optic primordia induces formation of duplicated retinas and extra lenses (Hyatt, et al., Proc. Natl. Acad. Sci. USA, 89:8293–8297, 1992). Patterning effects of retinoic acid upon the developing chick limb appear to be mediated through ectopic activation of the endogenous sonic hh gene (Riddle, et al., supra), these results with ectopic hh expression suggest the possibility of a similar mechanism underlying the patterning effects of retinoid acid treatment in the vertebrate eye.
To further elucidate the role of hh in eye development we utilized pax-2 and pax-6 (Krauss, et al, EMBO J, 10:3609–3619, 1991; Pitischel, et al., Development, 114:643–651, 1992) were utilized as positional markers to examine the effects of ectopic hh expression on the optic vesicle. As the optic vesicle evaginates from the lateral walls of the zebrafish forebrain (Schmitt, E. A. & Dowling, J. D., J. Comp. Neur., 344:532–542, 1994),pax-2 is expressed in a gradient, with highest RNA levels in the anterior and ventral regions of the optic vesicle (Krauss, et al., supra;
Superposition of developmental fate within the optic vesicle (Schmitt, et al., supra), upon the pattern of pax-2 expression suggests that the gradient of pax-2 RNA prefigures the future proximal/distal axis of the eye.
Ectopic hh alters the expression of pax-2, pax-6, and F-spondin. Zebrafish embryos were injected at the 1–8 cell stage with twhh or shh RNA and the pattern of pax-2, pax-6, or F-spondin expression was examined by whole mount in situ hybridization. Control embryos injected with lacZ RNA were performed in every case and displayed wild-type expression patterns. At embryo stage, the anterior-posterior axis of the optic vesicle corresponds to the future proximal-distal axis of the eye. During the next hour of development, the posterior edge of the optic vesicle will separate from the diencephalon (Schmitt and Dowling, Comp. Neur., 344:532–542, 1994).
Injection of either hh RNA causes uniform initiation of pax-2 expression along both the proximal-distal and dorsal-ventral axes of the optic vesicle as it begins to evaginate. The ectopic pax-2 expression appears at the same time as normal pax-2 expression is initiated in the eye, and in some cases, is also seen in the diencephalon between the optic vesicles. At the end of somitogenesis, a time when pax-2 would normally be restricted to the optic stalk, pax-2 RNA in hh injected embryos is detected in all but the most distal portion of the optic vesicle.
The effects of ectopic hh on expression of pax-6, which encodes a transcription factor critical for eye development was also studied. At 22 hours of zebrafish development, pax-6 is normally expressed in the lens and in most of the distal part of the optic cup (Krauss, et al., supra; Puschel, et al., Development, 114:643–651, 1992). In hh-injected embryos, pax-6 is repressed in the optic vesicle, although many embryos retain pax-6 expression in the most distal cells. With regard to pax-2 and pax-6 as markers of positional identity, hh expression in the optic vesicle can be characterized as inducing proximal fates and repressing distal fates.
The distal part of the optic vesicle is the most refractory to hh-induced changes in both pax-2 and pax-6 gene expression. Due to a later rotation, this distal portion of the optic vesicle will give rise to the dorsal portion of the mature eye (Schmitt, et al., supra); interestingly, this is the portion of the eye that remains in 3-day old injected embryos with intermediate phenotypes (see above).
Lesions in the pax-6 gene have been assigned as the basis for the Aniridia (Ton, et al., Cell, 67:1059–1074, 1991; Glaser, et al., Nat. Genetics, 2:232–239, 1992), Small eye (Hill, et al., Nature, 354:522–525, 1992), and eyeless mutations (Quiring, et al., Science 265:785–789, 1994), in humans, mice and Drosophila, respectively; pax-6 function thus appears to be critically required for eye development in Drosophila and mammals. As we argue here, hh-encoded activities also appear to play a role in vertebrate eye development, and this suggests a further molecular parallel between vertebrates and insects, since the role of hh in Drosophila eye development is well established (Mohler, et al., supra; Ma, et al., supra; Heberlein, et al., supra; Lee, et al., supra). The reciprocal and non-overlapping patterns of hh and pax-6 expression in the developing Drosophila eye (Ma, et al., supra; Quiring, et al., Science, 265:785–789, 1994), suggest the possibility of pax-6 repression by hh, but whether hh functions by similar mechanisms in vertebrate and Drosophila eye development is a questions that requires further investigation.
In mice, the dosage of pax-6 protein is crucial for normal eye development (Hill, et al., supra). Small eye heterozygotes develop an abnormally small lens (Hogan, et al., J. Embryol. Exp. Morph., 97:95–110, 1986; Hogan, et al., Development, 103 Suppl., 115–119, 1988), as do hh-injected embryos with weaker phenotypes (
The patterns of twhh and shh expression (
cycb16 (Hatta, et al., Nature, 350:339–341, 1991), heterozygous adults (a kind gift of R. Riggleman) were spawned and their offspring analyzed by whole mount in situ hybridization. Detection of pax-2 and either twhh or shh RNAs in embryos homozygous for the cyc mutation or their wild-type siblings. twhh RNA is only expressed in the presumptive tailbud (caret) of cyc embryos. As reported by Krauss, et al., Cell, supra, neural expression of shh is abolished in cyc embryos. Strong pax-2 expression was observed in the optic vesicles of wild-type embryos which is significantly reduced in cyc mutant embryos.
twhh RNA in cyclops embryos is found only in a small patch of cells at the presumptive tailbud and neural expression was not detected at any later stage examined. Neural expression of shh is also lost in eye mutants, although expression in the notochord is reunited (Krauss, et al., supra; data not shown).
Since the eye mutation appears to ablatc hh-expressing cells in the developing brain, this mutation can be used as a genetic tool to examine the requirement for hh function in eye development. Iiatta, et al.; Hatta, et al., Proc. Natl. Acad. Sci. USA, 91:2061–2065, 1994), recently demonstrated that pax-6 expression is fused at the midline due to loss of ventral midline cells that normally do not express pax-6 and, in addition, pax-2 expression in the fused eye of eye mutant embryos is reduced. We extended these observations to an earlier stage when the optic vesicles first form and found that pax-2 expression is weak and fails to extend within the vesicles in eye mutants. In conjunction with the results of ectopic hh expression, these observations suggest that hh signaling that activity promotes and is required for the induction of proxima fates within the eye vesicle. In this model, we propose that the protuberance acts as a proximal patterning center for the developing zebrafish eye by providing a localized source of hh activity.
Previous work has established an important role of signals from the floor plate and notochord in ventral patterning of the neural tube (Jessell, T. M., & Dodd, J., Cell, 69:95–110, 1992). For example, Goulding, et al., Development, 117:1001–1016, 1993, recently demonstrated that notochord and floor plate grafts can repress the normal lateral expression of pax-6 in the neural tube. Other recent work has implicated hh activity in at least some aspects of ventral neural tube patterning (Echelard, et al., Cell, 75:1417–1430, 1993; Krauss, et al., supra; Roelink, et al., supra); consequently, we examined hh-injected embryos for effects on pax-6 expression in the brain.
In the zebrafish at 22 hours of development, pax-6 is expressed in dorso-lateral regions of the diencephalon and in a ventro-lateral domain of the hindbrain and spinal cord that excludes the floor plate and adjacent cells (Krauss, et al., supra; Puschel, et al., supra). This pattern of expression is reciprocal to that of both twhh and shh in the diencephalon (compare
Since absence of pax-6 expression is a feature of the ventral midline, repression of pax-6 in lateral positions suggests ventralization. Consequently, twhh was injected into embryos for analysis of induction of a floor plate marker, F-spondin (Riddle, et al., supra). As described above, ectopic twhh induces F-spondin expression at more dorsal levels in the midbrain and anterior hindbrain. The effects of hh upon expression of both pax-6 and F-spondin indicate a ventralization of the brain. Adoption of ventral cell identity by lateral cells might explain their failure to form ventricles (
The ventralizing activities of twhh confirm and extend those previously reported for shh/vhh-1 class genes of chicken, zebrafish, and rat (Echelard, et al., supra; Krauss, et al., supra; Roelink et al., supra). The early restriction of twhh to midline neural progenitors, however, suggests that it may play a specific role in the homeogentic mechanisms of floor plate maintenance and expansion (Placzek, et al., Dev., 117:205–218, 1993). In the zebrafish, wild type cells in Cyclops hosts can contribute to and induce adjacent cells to form floor plate, but only when the transplanted cells populate the neural plate and not the notochord (Hatta, et al., Nature, 350:339–341, 1991). We have demonstrated that, in Cyclops mutants, midline expression of twhh is lost while shh expression is maintained in the notochord (
Endogenous hh protein in Drosophila is fund predominantly as an amino- and a carboxy-terminal fragment (N and C, respectively) derived by an internal auto-proteolytic cleavage of a larger precursor (U for uncleaved), which also occurs in vivo but at lower levels (Lee, et al., supra). Determinants within the amino-terminal domain appear not to be required for auto-proteolytic activity, whereas mutations affecting the carboxy-terminal domain can block auto-proteolysis and reduce activity in vivo (Lee, et al., supra). The auto-proteolysis is blocked by a substitution of alanine for the histidine normally present at position 329. This histidine is absolutely invariant in alignments of all known hh genes, and its sequence context suggests a catalytic role in auto-proteolysis (Lee, et al., supra).
The vertebrate hh proteins encoded by shh, twhh and mouse-shh/Hhg-1 also undergo auto-proteolysis to yield two smaller species from a single larger precursor (Lee, et al., supra; Chang, et al., supra; see lanes 1 and 6 in
Synthetic mRNAs transcribed from these constructs were injected to examine the role of processing and to assay the activities of individual protein fragments; the results are summarized in Table I and are based on the activities presented in
In considering the activity of delta N-C, it is important to recognize the activity of endogenous hh genes in these experiments, which are inhibited by delta N-C and fragments thereof (see Example 18 and
The uncleaved UHA protein is only somewhat less active than C in inducing pax-2, but it also was not able to repress pax6 efficiently (
In understanding the normal roles of N and C in eye and brain patterning, the N and C derivatives of the Drosophila hh gene may offer some insight. The Drosophila N derivative is retained close to its embryonic site of synthesis in a segmentally striped pattern (Tabata and Komberg, Cell, 76:89–102, 1994; Taylor, et al., Mech. Dev., 42 89–96, 1993), is cell-associated when expressed in cultured cells, and is effectively bound by heparin agarose in vitro, suggesting the possibility of extracellular matrix association. The C-terminal fragment, in contrast, is not bound effectively by heparin agarose, is almost quantitatively released into the culture supernatant of expressing cultured cells, and is only diffusely localized in embryos. Although the activities of individual fragments have not been assayed, the biochemical differences and tissue distributions of Drosophila N and C may account for the short and long range nature of the functions associated with hh during Drosophila development.
Although the tissue distributions of zebrafish N and C are not known, their activities in ectopic expression assays are also suggestive of short- and long-range functions when considered in the context of normal expression patterns of hh, pax-2 and pax-6. The normal gradient of pax-2 expression in the optic vesicle extends a substantial distance from its maximum adjacent to the site of hh expression in the protuberance; the ability of ectopic C to activate pax-2 therefore suggests that, consistent with the distribution of C in Drosophila, zebrafish C may carry out a long-range function. Repression of endogenous pax-6 expression, in contrast, appears to be a short-range function since pax-6 expression occurs close to endogenous hh expression. Efficient repression of pax-6 is an attribute of constructs producing N, and a short-range function for N would be consistent with the distribution of N in Drosophila.
Two types of hh-dependent activity have been reported for hh-transfected cultured cells. One is the apparent contact-dependent induction of floor plate markers (Roelink, H., et al., Cell 76:761–775, 1994); the second induction of sclerotome markers in presomitic mesoderm, is diffusible and acts at long-range.
1. Materials and Methods
cDNAs encoding full-length Xenopus hedgehogs, or encoding amino terminal or carboxy terminal domains linked to secretory leader sequences were transcribed in vitro to yield translatable messenger RNA. The synthetic messenger RNAs, and control mRNAs, were microinjected into the animal poles of cleavage stage Xenopus embryos, which were allowed to develop to the blastula stage, at which time the animal cap explants were prepared from the upper one fourth of the embryo. These blastula cap explants were then cultured in vitro in physiological saline in the presence or absence of the transforming growth factor beta family member, recombinant human activin A. All explants were allowed to develop until control embryos had grown to neurula stage, or to tadpole stage. Importantly, blastula caps left untreated differentiate from ectoderm into a typical epidermis. Blastula caps treated with activin differentiate into mesodermal and neural cell types. Thus, the question was whether hedgehog, or its proteolytic derivatives, would change the differentiation of cells away from becoming epidermis, and into another cell type. A second question was whether hedgehog can work with activin to alter the normal response of the tissue to either factor by itself.
Explants were then extracted to yield mRNA by methods commonly used by those of skill in the art, which was used as template with reverse transcriptase to yield cDNA. The cDNA was then used as template with various sets of primers for PCR for specific genes, reverse-transcriptase-polymerase chain reaction, or RT-PCR. This results in specific amplification of radioactive products which are diagnostic for the presence and level of the messenger RNAs which were present in the explants. Samples were separated on polyacrylamide gels, which were exposed to X-ray film to yield the bands shown in the figures. Thus, the darker bands correspond to a greater level of the specific mRNA.
The panel labelled XANF1B detects a pituitary gene. Lane 4 (panel A) shows that hedgehog induces this pituitary marker, and thus likely pituitary cell types, in blastula cap explants (see also
The panel labelled OTX-A detects this anterior brain gene. Lane 4 (and 6 in
In panel 18B, embryos were injected with N or ΔN-C, and some animal cap explants were treated with activin before culturing until sibling embryos reached tailbud stage. Lanes 1, 2: control animal caps from uninjected embryos. Lanes 3, 4: control animal caps from uninjected embryos, treated with activin. Lanes 5, 6: animal caps from embryos injected with N and treated with activin. Lanes 7, 8: animal caps from embryos injected with ΔN-C and treated With activin. Whereas N displays activities in activin-treated explants similar to those of X-bhh (see B) ΔN-C produces the opposite effect, decreasing anterior and increasing posterior neural marker expression. As shown in
Examining the first panel for the cement gland marker XAG-1 clearly shows that intact hedgehog (lane 6) and the N domain (lane 8) and the processing defective hedgehog (lane 12) are much better than inducing the cement gland than is the C domain (lane 1). Examining the second panel demonstrates that the C domain (lane 10) is better at inducing the pituitary gene XANF1B than is the N domain (lane 8). Since the N domain induces the XAG-1 marker better, described in point A above, the two results together clearly demonstrate that the N and C domains have distinguishable activities. Examination of the remaining panels shows that all described activities of the normal hedgehog (lane 6) can be defined in terms of the activities of the N and C domain.
Examining the third panel, for the forebrain gene otx-A, shows that both the N domain (lane 8) and C domain (lane 10) induce similar levels of this gene, but the processing defective hedgehog (lane 12) is better than either at inducing this gene.
Examining the fourth panel of this figure (NCAM), (as well as the
Experiments in the Xenopus embryo were conducted by injecting full-length hedgehog RNA, and immunoprecipitating with a C-domain specific antibody, which proves that full length hedgehog does in fact get processed in vivo in vertebrates, consistent with the data shown in earlier Examples in Drosophila. Thus, the ideas for the utility of detecting hedgehog N and C domains is based on knowledge that such domains do appear through hedgehog processing in vertebrates. Moreover, the knowledge that hedgehog processing does occur in vivo naturally raised the question of whether the resulting N and C domains have independent activity.
The results in
The data in
An internal deletion of X-bhh (ΔN-C) blocked the activity of X-bhh and N in explants and reduced dorsoanterior structures in embryos. As elevated hh activity increases the expression of anterior neural genes, and as ΔN-C reduces dorsoanterior structures, these complementary data support a role for hh in neural induction and anteroposterior patterning.
ΔN-C deletes amino acids 28–194 of X-bhh. The primary translation product is predicted to undergo signal sequence cleavage removing amino acids 1–23, and to undergo autoproteolysis. Based on the cleavage site in Drosophila hh (Porter, et al., Nature, 374:363, 1995) autoproteolysis would generate a C domain of X-bhh amino acids 198–409, as well as a predicted seven amino acid polypeptide, representing amino acids 24–27, and 195–197 (Lai, et al., Development 121:2349, 1995). Analysis of the effect of ΔN-C on neural markers was by standard methods including Northern blot analysis and in situ hybridization (Lai, et al., supra, incorporated herein by reference).
Although ΔN-C does not induce the cement gland marker XAG-1, it decreases the expression of anterior ectodermal and neural markers in activin-treated animal caps. Thus, ΔN-C has the capacity to affect neural patterning. ΔN-C also promotes an increase in posterior neural markers in activin-treated animal caps. Mixing ΔN-C with N or full length X-bhh at a 1:1 ratio led to a dramatic inhibition of the induction of cement gland in animal cap assays, supporting the hypothesis that ΔN-C interfered with X-hh.
In addition to peptide bond cleavage, Hh autoprocessing causes the covalent attachment of a lipophilic adduct to the COOH-terminus of Hh-Np (J. A.Porter et al., Cell 86, 21, 1996). This modification is critical for the spatially restricted tissue localization of the Hh signal; in its absence, the signaling domain exerts an inappropriate influence beyond its site of expression (J. A. Porter et al., Cell 86, 21, 1996). Physical and biochemical characterization of this lipophilic adduct indicates that it is not the glycosyl phosphatidyl inositol (GPI) anchor, the only other known lipophilic modification associated with secreted cell surface proteins in eukaryotes (S. Udenfriend and K. Kodukula, Annu. Rev. Biochem. 64, 563, 1995; and P. J. Casey, Science 268, 221, 1995).
In vitro studies of Hh autoprocessing were performed using a bacterially expressed derivative of the Drosophila protein, His6Hh-C, in which the majority of the NH2-terminal signaling domain and the signal sequence are replaced by a hexa-histidine tag. Cleavage of this protein occurs between residues corresponding to Gly 257 and Cys 258 (J. A. Porter et al., Nature 374, 363, 1995) and likely proceeds through a labile thioester intermediate formed by the cysteine thiol and the glycine carbonyl carbon. In the presence of high concentrations of thiols or other small molecules with strongly nucleophilic properties at neutral pH, cleavage of the peptide results from nucleophilic attack upon the thioester carbonyl, causing displacement of the thiol group and formation of an adduct to Gly 257 by the attacking nucleophilic (
The in vivo reaction resulted in lipophilic modification of the NH2-terminal signaling domain. The most direct mechanism by which this could occur, by analogy to the in vitro mechanism (
To identify the components active in the reaction, the bulk S2 lipids were separated into two classes, neutral and complex, by silicic acid column chromatography (W. W. Christie, Lipid analysis (Pergamon, Oxford, ed. 2nd, 1982).
The activity was found exclusively in the neutral class, so the lipids were subjected to preparative thin layer chromatography (TLC) using a solvent system that resolved neutral lipids (W. W. Christie, Lipid analysis (Pergamon, Oxford, ed. 2nd, 1982) (
With the use of various lipid standards, it was found that spot B comigrated with cholesterol (
The part of the sterol most likely to act as attacking is the 3β hydroxyl. Such an attack would leave cholesterol as a covalent adduct in ester linkage to the carboxylate of the terminal residue of the NH2-terminal fragment (GLY257).
Ester bonds are subject to hydrolysis in alkaline conditions and base treatment prior to electrophoresis indeed reduced the migration of the cholesterol-driven reaction product to a position coinciding with that of the dithiothreitol-driven reaction product. These results are consistent with stimulation of the in vitro processing reaction by direct nucleophilic attack of cholesterol on the thioester intermediate to form an ester-linked adduct. If processing of Hh also results in formation of an ester-linked cholesterol adduct in vivo, then the protein-lipid linkage should be subject to base hydrolysis with a concomitant shift in electrophoretic mobility of the protein (normally 18.5 kD). The immunoblot in
To confirm the involvement of cholesterol in formation of the Hh-Np adduct in vivo, S2 cells containing an inducible wild-type Hh construct were metabolically labeled with [3H]cholesterol.
After 48 hours of growth in the presence of [3H]cholesterol, induced and uninduced cultured cells were detergent extracted and total cell proteins were subjected to SDS-PAGE followed by fluorography (Metabolic labeling of S2 cultured cells with [3H]cholesterol was performed essentially as described (Silberkang, et al., J Biol. Chem. 258:8503, 1983). Briefly, cells containing a stably integrated Cu++-inducible hedgehog gene were grown at 23° C. for two weeks in Schneider cell media (Gibco) containing a 5% fetal bovine serum depleted of lipoprotein (low cholesterol media, ˜20 μg/ml cholesterol). These cells were then plated at 40% confluence onto two 35 mm tissue culture dishes (Nunc) in 1 ml of low cholesterol media supplemented with 300 μCi of labeled cholesterol, [18.104.22.168-3H(N)] 65 Ci/mM (NEN) giving a specific activity for cholesterol in this medium of ˜5 Ci/mmol. After 24 hours (1 doubling time) one plate of cells was induced to express Hh protein by the addition of CuSO4 (1 mM final concentration). After an additional 24 hours the cells from both dishes were harvested, lysed in Tris buffered saline containing 1% Triton X-100, and total cell protein was precipitated with 5 volumes of cold acetone. The protein pellet was resuspended in 2% SDS in H2O and reprecipitated with acetone several times to remove unincorporated radioactivity prior to loading onto SDS polyacrylamide gels for analysis. Initial labeling experiments in which 25 μCi of cholesterol was added resulted in ˜10 fold decrease in extent of label incorporated into the inducible Hh-Np protein). Whereas uninduced cells showed no incorporation of [3H]cholesterol into cellular proteins, cells induced to express Hh showed a single strong band with a mobility corresponding to that of Hh-Np (
The amount of radioactive cholesterol incorporated is consistent with that expected if all of the Hh-Np synthesized upon induction received a cholesterol adduct (The specific activity of [3H]cholesterol in the S2 cell labeling medium was ˜5 Ci/mmol. Assuming after a 24 hour doubling time that this concentration approximately represents that within the S2 cell membrane, then any protein subsequently expressed and receiving cholesterol as an adduct would also be labeled at the same specific activity. As determined by standardized coomassie blue staining, −50–100 ng or 2.5 to 5 picomoles of Hh-Np is produced by one 35 mm dish of S2 cells containing the Cu++-inducible Hh construct during 24 hours of induction with 1 mM CuSO4 (13). This predicts ˜12.5 to 25 nCi or 2.75×104 to 5.5×104 dpm of radioactivity would be incorporated into Hh-Np protein produced in our labeling experiment assuming it is cholesterol modified. Total incorporation of radioactivity into Hh-Np during the in vivo labeling experiment described above was measured at ˜5×104 dpm by excision and scintillation counting of an Hh-Np gel band), suggesting that other cellular components do not complete effectively as nucleophilic adducts in the in vivo autoprocessing reaction. Also consistent with a homogenous adduct, the mass of cholesterol is consistent with the mass previously measured by mass spectrometry of processed protein purified from cultured cells. A recent MALDI mass spectral analysis gave a mass of ˜430 daltons for the Hh-Np adduct, ˜9% larger than the mass of cholesterol (386.6). Detection of this modification required that Hh-Np be treated with CNBr/70% formic acid, i.e. full length Hh-Np could not be detected. The mass discrepancy noted above could be accounted for by the net addition of formic acid (45 daltons) during CNBr digestion. This reaction could involve the addition of H2O across the 5,6 double bond of cholesterol, a common reaction of secondary alkenes in strong acids [R. T. Morrison, R. N. Boyd, Organic Chemistry (Allyn and Bacon, Boston, ed. 3rd, 1973)], followed by esterification of formate via this newly formed alcohol [B. I. Cohen, G. S. Tint, T. Kuramoto, E. H. Mosbach, Steroids 25, 365–378, 1975. To test whether the sterol backbone could be modified by the CNBr treatment, a positively charged cholesterol derivative (3β(N′,N′-dimethylamino) ethanecarbamoyl)-cholesterol, Sigma) detectable by MALDI was examined. It was found that incubation of this sterol derivative in 70% formic acid alone resulted in the addition of 45 mass units to the sterol (13), a mass consistent with the net addition of a formic acid molecule). These in vitro and in vivo results show that the Hh-C processing domain functions as a cholesterol transferase; as a result of this activity, a cholesterol adduct is attached via an ester linkage to the COOH-terminus of the NH2-terminal signaling domain of the Hh protein.
To test whether processing of vertebrate hedgehog proteins results in the incorporation of cholesterol as a covalent adduct to the signaling domain, cultured green monkey kidney cells (COS-7) were metabolically labeled with [3H]cholesterol and transfected with expression constructs containing (i) the full length murine Sonic hedgehog (Shh) open reading frame, leading to production of an autocatalytically processed signaling domain (Shh-Np) or (ii) Shh coding sequences precisely truncated at the site of cleavage, thus producing an unprocessed amino terminal signaling domain (Shh-N) (COS-7 cells grown at 37° C. in DMEM supplemented with 10% fetal calf serum were plated at ˜35% confluence onto two 35 mm dishes in 1 ml of Optimem media (Gibco) containing 1.5% fetal bovine sera and 25 μCi of cholesterol, giving ˜40 μg/ml as the final concentration of cholesterol with a specific activity of 2 Ci/mmol (labeling medium). After 24 hours the labeling medium was removed and the cells were transfected for 6 hours with Shh or Shh-N expression constructs using lipofectamine (Gibco) and serum-free DMEM media. After transfection, 1 ml of fresh labeling medium was added to each dish and the cells were incubated for 36 hours at 37° C. The cells were then harvested without washing, lysed on the plate with Tris buffered saline plus 1% Triton X-100 and the total cell proteins were precipitated with acetone, washed and analyzed as described above for the S2 cell proteins). Cells expressing the full length construct contained a prominent radio-labeled species migrating at ˜19 kD, suggesting that cholesterol is covalently added to Shh-Np (
An experimental model for holoprosencephaly derives from the occurrence of epidemics of congenital craniofacial malformations among newborn lambs on sheep ranches in several National Forests of the western United States. The most dramatically affected lambs showed severe holoprosencephaly, including true cyclopia and other craniofacial malformations characteristic of holoprosencephaly. The occurrence of these defects was traced to grazing by pregnant ewes on the range plant Veratrum californicum. The compounds responsible were identified as a family of steroidal alkaloids; the structures of two of these, cyclopamine and jervine, are shown as compared to cholesterol in
Given the structural similarities of these compounds to cholesterol and the similar teratogenic effects of cholesterol synthesis inhibitors upon the offspring of pregnant rats, a reasonable mechanism to consider for the effects of these plant sterol derivatives was the inhibition of cholesterol biosynthesis. Accordingly, COS7 cultured cells treated with jervine were tested for defects in cholesterol biosynthesis by labelling with [3H]-mevalonic acid and then extracting and analyzing radiolabelled, non-saponifiable lipids.
Metabolic labeling and sterol analysis was essentially as described (Popiak et al. J. Biol. Chem. 264: 630–6238.1989; Rilling et al. 1993 Arch. Biochem. Biophys. 301: 210–215.), with minor modifications. Briefly, COS-7 cells were plated at ˜35% confluence into two 60 mm dishes at 37° C. in 4 ml each of Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS). After 24 hr of growth the medium in each dish was replaced with 2 ml fresh medium with 10% FBS; [3H]-mevalonic acid (NEN #NET 176) brought to a specific activity of 0.8 Ci/mmol in a 1% solution of bovine serum albumin was added to this medium to a final concentration of 20 mM. At this time, one dish received 6 ml of a 4 mg/ml solution of jervine in ethanol (final concentration 28 mM jervine), and the other received 6 ml of ethanol. After 24 hr further incubation, cells were washed in PBS, extracted with methanol, and 1 M potassium hydroxide (KOH) added to 10%. Following a three hour incubation at 60° C., the methanol/KOH mixture was extracted with diethyl ether, the extract dried down, resuspended in isopropanol, and subjected to reverse phase HPLC analysis by the method of Rodriguez and Parks (Methods in Enzymology 111: 37–511985).
Treated cells synthesized reduced levels of cholesterol and accumulated increased levels of another sterol that we have provisionally identified as the cholesterol precursor, zymosterol. The natural product jervine at these concentrations thus inhibits cholesterol biosynthesis in cultured cells in much the same manner as the synthetic drugs discussed above, although the specific enzyme(s) affected appear to differ. Given the similarities in their teratogenic effects, this inhibition seems likely to underlie the teratogenic effects of both the synthetic and natural compounds.
Protein expression and purification: Drosophila melanogaster Hh protein in which most of the amino-terminal signaling domain and signal sequence have been replaced by a hexa-histidine tag (His6Hh-C25) was expressed as previously described (Porter et al., 1995). SeMet His6Hh-C25 was prepared by expression in E. coli strain B834 (DE3) pLysS, a methionine auxotroph, and growth in minimal media as previously described (Leahy et al., 1994). This His6-tagged protein was purified on a Ni++-NTA agarose column arid autocleavage stimulated by addition of 50 mM DTT. After removal of the DTT by dialysis, the cleaved protein was passed over a Ni++-NTA agarose column and the Hh carboxy-terminal domain, Hh-C25, collected in the column run through. Hh-C25 was subjected co limited proteolysis by overnight incubation with 1:500 (w:w) subtilisin (Boehringer Mannheim) at 4° C. A protease-stable fragment or approximately 17 kDa, Hh-C17, was identified by SDS-PAGE and purified by anion-exchange chromatography utilizing a Mono-Q column (Pharmacia). The amino- and carboxy-terminal residues of Hh-C17 were determined to be Cys-258 and Ser408, respectively, by mass spectral analysis of cyanogen bromide-cleaved fragments. Mass spectral analysis was performed as previously described (Porter et al., 1996a)
Crystallization: Crystals were grown from hanging drops by the method of vapor diffusion (Wlodawer et al., 1975). 6 μl of a 1.4 mg/ml solution of Hh-C17 in 1.4 mM β-mercaptoethanol were mixed with 2 μl of a 1:1 dilution of reservoir solution (20% PEG 3350, 80 mM ammonium sulfate, and 10 mM sodium cacodylate, pH 5.8) with distilled water and equilibrated over the reservoir solution. Crystals typically grew to a final size of 0.2 mm×0.2 mm×0.1 mm over 3–7 days. Crystals are in space group I2t3 with unit cell dimension a=b=c=101.54 Å.
Data collection and processing: All data were collected from crystals soaked in mother liquor made 10% (w/v) ethylene glycol and flash frozen in a gaseous nitrogen stream at −180° C. MAD data were collected at four wavelengths from a single SeMet crystal at beamline X-4A of the National Synchrotron Light Source at Brookhaven National Laboratory. Data were collected using Fuji HIR-V phosphor-imaging plates and digitized using a Fuji BA-3000 scanner. 2° oscillations at φ and φ+180° were collected with no overlap for each oscillation range at each wavelength. All diffraction images were processed using the program DENZO and scaled with the program SCALEPACK (Orwinowski and Minor, 1997). <I+> and <I−> were used for MAD phase determination and partially recorded reflections were used in all cases. Diffraction data from different wavelengths were scaled with WVLSCL, and values for FA and optimal f′ and f″ were calculated with MADLSQ (Hendrickson, 1991). Data collection statistics are shown in Table 1.
Structure determination: Three selenium sites were deduced from FA amplitudes using both the program SHELXS (Sheldrick, 1991) and Patterson methods. MAD phase determinations were made with the program MLPHARE (Collaborative Computational Project, 1994; Ramakrishnan and Biou. 1997), and solvent-flattening and histogram-matching were performed with the program DM (Collaborative Computational Project, 1994). An atomic model consisting of Cys-258 to Tyr-401 was readily built into electron density maps computed with MAD-derived phases for reflections in the range 20.0–2.0 Å using the program “O” (Jones et al., 1991). One round of simulated annealing and several rounds of Powell minimization using X-PLOR (Brënger, 1992) alternated with model building with “O” yielded the current model of Hh-C17 consisting of 145 residues, Cys-258 to Ala-402, and 126 water molecules. The model was refined using the data collected at 0.9919 Å. One molecule is present in the asymmetric unit, and the solvent content is approximately 59%. All backbone torsion angles are within energetically acceptable regions. No electron density was observed for residues 403 to 408, but additional electron density was observed near the thiol group of Cys-258. As the crystallization buffer contained cacodylic acid, both AsO(CH3)2 or an As atom were modeled in this density, but neither the crystallographic R-factor nor the free R-factor improved with these atoms added to the refinement and no atoms have been included in this region in the final atomic model.
Site-directed mutagenesis and in vitro autocleavage assays: His-329, Thr-326, and Asp-303 were each mutated to alanine (H329A, T326A, and D303A, respectively) and Leu-409 was mutated to a stop codon (His6Hh-C17) by the method of recombinant circle PCR (Jones and Winistorfer, 1992). His6-tagged proteins containing residues 83–471 of Drosophila Hh protein with residues 89–254 deleted and with the mutated residues were expressed in E. coli and purified to near homogeneity as previously described (Porter et al. 1995). Autocleavage activity of the mutant proteins was assessed in 15 μl reactions by incubating 1 μg of protein in 150 mM NaCl, 100 mM Tris-HCl (pH 7.4), 0.05% Triton X-100, 1.25 mM β-mercaptoethanol, and 2.5% glycerol with either 50 mM DTT or 350 μM cholesterol/1 mM DTT for 6 hours at 30° C. The cleavage products were then fractionated by SDS-PAGE and detected by Coomassie Brilliant Blue staining. The activity of D303A and His6Hh-C17 were also assessed by incubating 1 μg of protein with 46 μM [3H]cholesterol (11.6 Ci/mmole)/1 mM DTT for 6 hours at 30° C. The proteins were then subjected to SDS-PAGE and labeled proteins detected by autoradiography.
Database searching and sequence alignment: Screening of the non-redundant protein sequence database at the National Center for Biotechnology Information (NIH) was performed using the BLASTPGP program, which is an enhanced version of BLAST that produces gapped alignments (Altschul and Gish, 1996). Additional searches were Performed using the PSI-BLAST (Position-Specific Iterative BLAST) program, which constructs position-specific weight matrices from the BLASTPGP output and employs them for subsequent iterations of database screening using a modification of the BLAST statistics (Altschul et al., in press). Alignments of multiple protein sequences were constructed using the CLUSTALW program (Thompson et al., 1994) or the MACAW program (Schuler et al., 1991).
Domain Identification and Structure Determination
Drosophila melanogaster Hh in which the signal sequence and most of the amino-terminal signaling domain have been replaced by a hexahistidine tag was expressed in E. coli as previously described (Porter et al., 1995). Following purification with Ni++-NTA agarose, this protein cleaves itself in vitro in the presence of either DTT or cholesterol to liberate the 25 kDa Hh-C fragment (Hh-C25, residues Cys-258 to Asp-471) Hh-C25 prepared by this method was found to be poorly soluble in the absence of detergents and susceptible to further proteolytic breakdown when concentrated to 1 mg/ml or greater. Treatment of Hh-C25 with subtilisin, however, resulted in a protease-stable fragment of ˜17 kDa molecular weight (Hh-C17) with improved solubility. Mass spectrometric analysis of cyanogen bromide cleavage fragments of Hh-C17 showed it to consist of residues Cys-258 to Ser-408 (data not shown). All residues absolutely conserved in Hh-C homologues (Porter et al., 1996a), including the nematode sequences, are contained in Hh-C17. To determine if Hh-C17 retained autoprocessing activity, a mutant version of His-tagged Hh-C containing a termination codon at residue position 409 (His61Hh-C17) was expressed and assayed for autocleavage in the presence of DTT and cholesterol. As shown in
Crystals of Hh-C17 that diffracted to at least 1.9 Å Bragg spacings were readily produced from both native and selenomethionyl-substituted (SeMet) protein. The crystal structure of Hh-C17 was determined by the method of multiwavelength anomalous diffraction (MAD) usin SeMet crystals (Hendrickson et al. 1990; Hendrickson, 1991). High quality experimental electron density maps allowed construction of an atomic model for Hh-C17 residues Cys-258 to Ala-402 that readily refined to low R-factor with good stereochemistry. Final refinement and stereochemical statistics are summarized in Table 1.
Description of Hh-C17 Structure
Hh-C17 possesses an all-β structure that is roughly disk-shaped with a diameter of −35 Å and width of −20 Å. The amino and carboxy termini emerge from the same surface of Hh-C17˜6 Å apart. A ribbon drawing and topology diagram of the Hh-C17 structure are shown in
Despite a low level of sequence conservation, the two Drosophila Hh-C17 subdomains are superimposable with an r.m.s deviation in α-carbon positions of 1.38 Å, and several notable structural features, including β-bulges and specific β-turn types, are conserved between the subdomains (
The level of structural similarity between the two Hh-C17 subdomains suggests that Hh-C17 could have arisen by tandem duplication of a primordial gene. The duplicated sequences do not, however, correspond directly to the compact subdomains observed in the Hh-C17 structure. As can be seen in
The exchange of domains or elements of secondary structure has been observed in several proteins and is believed to result in a more stable association of subunits in multidomain proteins (Bennett et al., 1995). Exchange of structural regions has principally been observed between independent polypeptide chains within homodimers, but the lac operon repressor and homologues also appear to represent a case of exchange between duplicated domains within a single polypeptide chain (Schumacher et al., 1994; Lewis et al., 1996).
Active Site Residues
The amino-terminal residue of Hh-C17, Cys-258, is involved in both the thioester formation and cholesterol transfer steps of Hh autoprocessing (see
To assess the involvement of His-329, Thz-326, and Asp-303 in Hh autoprocessing7 each of these residues was mutated to alanine within the context of the full-length His6Hh-C25 protein and the mutant proteins were expressed and assayed for Hh autoprocessing activity. The autocleaving activity of the mutant proteins in the presence of high concentrations or DTT was used as an assay for thioester formation, the first step in the Hh autoprocessing reaction, while the autocleaving activity in the presence of cholesterol was used to assay for cholesterol transfer, the second step in the autoprocessing reaction. The results of these assays are shown in
The loss or dramatic reduction of autocleaving activity in the presence of both DTT and cholesterol for H329A and T326A implicates both His-329 and Thr-326 in formation of the internal thioester during Hh autoprocessing. The interaction of the side chains of both of these residues with the α-amino group of Cys-258, a component of the cleaved peptide bond, strongly implies a direct role for these residues in thioester formation. Possible roles for His-329 during thioester formation include stabilization of negative charge on the carbonyl oxygen of Gly-257, donation of a proton to the free α-amino group of Cys-258, and maintenance of an appropriate orientation of reaction components through polar interactions. His-329 may also deprotonate the thiol group of Cys-258 prior to thioester formation, but if this is the case some rearrangement of Cys-258 relative to its position in the Hh-C17 crystal structure would be required to bring the thiol group of Cys-258 into proximity with His-329. As the pKa of the thiol group in free cysteine is 8.3, a base may not be needed to catalyze thiol deprotonation. Possible roles for Thr-326 in thioester formation seem more limited. The high pKa of a threonine hydroxyl group (>15) makes Thr-326 an unlikely candidate for proton transfers, suggesting that this residue is needed to form polar interactions that stabilize reactive conformations within the Hh protein.
The activity of the D303A mutant in DTT—but not cholesterol-stimulated autoprocessing shows that Asp-303 is not needed for thioester formation but is required for cholesterol transfer. The negatively-charged aspartic acid residue seems unlikely to be involved in binding a hydrophobic cholesterol molecule. A role in activating the cholesterol molecule for nucleophilic attack of the thioester appears more plausible. For cholesterol to become an effective nucleophile, the 3β-hydroxyl group must become deprotonated, and Asp-303 is a good candidate for the general base that catalyzes this deprotonation. Substitution of Asp -303 with histidine in Hh-C homologues is consistent with this hypothesis as histidine is also capable of functioning as a general base.
As indicated by the inactivity of Hh-C17 in cholesterol transfer assays, residues in the 63 amino acids removed from the Hh-C25 carboxy terminus are also involved in cholesterol transfer. residues in cholesterol binding or activation. The decreased solubility of Hh-C25 relative to Hh-C17 suggests that the carboxy-terminal 63 residues of Hh-C25 may possess an exposed hydrophobic region that could serve as a cholesterol binding site.
Relationship between Hh-C17 and Self-Splicing Proteins
An earlier analysis identified a 36 amino acid conserved motif in the amino-terminal regions of Hh-C homologues and inteins (Koonin, 1995). A greatly expanded database of Hh-C and intein sequences coupled with recent enhancements of the BLAST method for database searching enabled extension of the detectable region of sequence similarity to the amino-terminal ˜100 amino acids of Hh-C and intein sequences (p ˜10−3–10 −4). The improved methods for database searching include statistical analysis of gapped alignments and iterative database scanning with position-specific matrices derived from previous BLAST outputs (Altschul et al., in press). When a database search was initiated with any of the Hh-C sequences or with most of the intein sequences, members of the respective second protein family were the only additional sequences retrieved from the database at a statistically significant level.
Solution of the Hh-C17 crystal structure showed the expanded region of Hh-C/intein sequence homology to terminate halfway through one of the subdomains in the turn region of an exposed loop between β strands 3 b and 4 b (see
As can be seen in
The only residue absolutely conserved between Hh-C homologues and inteins is a histidine corresponding to His-329 in Drosophila Hh-C. The presence of His-329 in the active site of Drosophila Hh-C and the loss of thioester formation activity when His-329 is mutated strongly imply that this histidine is conserved because it performs a vital role in thioester formation and that it functions similarly in inteins and Hh-C homologues. The only other residue conserved in the active site of Hh-C homologues and shown by mutagenesis to be required for efficient thioester formation, Thr-326, is also extremely conserved in intein sequences. Of the 39 intein sequences in the database at the time of our comparison, 34 sequences contain a threonine at a homologous position to Thr-326, While three inteins have serine, and one each have asparagine or glutamic acid at this position. The high level of conservation of threonine at this active site position and its substitution with similar amino acids suggests a conserved role for this threonine in inteins and Hh-C homologues. A conserved residue homologous to Asp-303, also found in the Hh-C17 active site, is not found in intein sequences, consistent with its role in cholesterol activation rather than thioester formation.
As expected from the sequence homology, die structures of the self-splicing region of the PI-SceI intein (Duan et al., 1997) and Hh-C17 are clearly homologous. Although not previously noted, the self-splicing region of PI-SceI contains homologous subdomains related by pseudosymmetry. The PI-SceI subdomains are homologous to the Hh-C17 subdomains and possess the same loop exchange observed in Hh-C17. However, these features are obscured by insertion of endonuclease-associated sequences. In addition to insertion of the core endonuclease domain in the region homologous to the β3b-β4b loop, the PI-SceI intein contains an additional insertion of amino acids relative to the Hh-C17 structure. The site of this insertion occurs in the turn between β strands 1b and 2b in the Hh-C1-7 structure (see
The conservation of structure, sequence and cleavage mechanism between Hh-C homologues and the intein regions of self-splicing proteins firmly establishes the divergence of these two protein families from a common precursor.
Table I. Statistics for data Collection, Phase Determination and Refinement
(A) Rsym and completeness values were calculated considering Bijvoets equivalent. Values in parentheses for <I/σI> are for the highest resolution shell (1.98–1.9 Å). Rsym=100×ΣhΣi|Ii(h) −<I(h)>|/ΣhΣiIi(h). (B) r.m.s. (Δ|F℄)/r.m.s. (|F|) where ΔF is the Bijvoet difference at one wavelength (values on the diagonal) or the dispersive difference between two wavelengths (values off the diagonal). Also shown are the anomalous components of the Se scattering factors as a function of wavelength as determined by MADLSQ (Hendrickson, 1991). (C) All data for which |F|2σ were used in the refinement. A subset of the data (10%) was excluded from the refinement and used to calculate the free R-value (Brünger, 1992). A final round of refinement including this data was performed to produce the final set of coordinates and crystallographic R-value R-value=Σ∥Fo|−|Fc∥/Σ|Fo|.
The above disclosure generally describes the present invention. A more complete understanding can be obtained by reference to the following specific examples which are provided herein for purposes of illustration only and are not intended to limit the scope of the invention.